- Email: [email protected]
Surgical repair of the rotator cuff: a biomechanical evaluation of different tendon grasping and bone suture fixation techniques
Clinical Biomechanics 18 (2003) 721–729 www.elsevier.com/locate/clinbiomech
Surgical repair of the rotator cuff: a biomechanical evaluation of different tendon grasping and bone suture fixation techniques Massimiliano Baleani a
a,*
, Stephan Schrader a, Carlo Andrea Veronesi b, Roberto Rotini b, Roberto Giardino c, Aldo Toni a,d
Laboratorio di Tecnologia Medica, Istituti Ortopedici Rizzoli, Via Di Barbiano 1/10, 40136 Bologna, Italy b Modulo Dipartimentale Chirurgia Spalla-Gomito, Istituti Ortopedici Rizzoli, 40136 Bologna, Italy c Servizio di Chirurgia Sperimentale, Istituti Ortopedici Rizzoli, 40136 Bologna, Italy d I Divisione Ortopedia e Traumatologia, Istituti Ortopedici Rizzoli, 40136 Bologna, Italy Received 28 January 2002; accepted 28 May 2003
Abstract Objective. This study investigated the initial strength and failure mode of different rotator cuff repair techniques. Background. Full or partial re-rupture of the repair is one of the main post-operative complications for rotator cuff repair. The rate of failure is strongly affected by the extension of the tear, increasing in case of large or massive tears up to 62%. Design. The study was planned to assess the three individual components of the tendon-to-bone repair (tendon grasping, suture knotting, suture-to-bone fixation) and to identify the best combinations in terms of mechanical strength to failure. The best combinations were tested to compare the mechanical behaviour of the entire repair and suggest potential improvements in the repair technique. Methods. Experimental tests were performed using sheep shoulders. Three tendon-grasping techniques, two suture knotting techniques, and the effect of bone augmentation with metallic plate and bone quality on suture-to-bone fixation were investigated. Results. This study assessed the mechanical behaviour of different repair components. The best combinations of the investigated techniques showed that the weakest link was the tendon–suture interface. More importantly, the compliance of the investigated repairs was large. Conclusions. The initial strength of the rotator cuff repair can be improved by changing the repair technique. Nevertheless, even a low physiological load stressing the repaired tendon may cause a gap formation at the tendon–bone interface without necessarily producing failure of the repair. Relevance Post-operative protection of the repaired rotator cuff from tension load is necessary to reduce the risk of delaying or preventing of the healing process. 2003 Elsevier Ltd. All rights reserved. Keywords: Rotator cuff; Repair technique; Biomechanics; Experimental testing; Tensile strength; Stiffness
1. Introduction Lesion of the rotator cuff is a common problem in orthopaedics. Progressive pain or functional deficits to the shoulder joint generally requires surgical repair of the rotator cuff. Traditional techniques for the repair of the rotator cuff have evolved from the work of
*
Corresponding author. E-mail address: [email protected] (M. Baleani).
0268-0033/$ - see front matter 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0268-0033(03)00122-0
McLaughlin (1944). McLaughlin described a surgical method for the reattachment of the avulsed rotator cuff tendon to the humeral bone by passing sutures through transosseous tunnels in the greater tuberosity. Although others have subsequently studied the use of different fixation techniques, trying to obtain a greater strength of fixation or to define a less invasive technique (Craft et al., 1996; France et al., 1989; Goradia et al., 2001; Hecker et al., 1993; Gerber et al., 1994; Reed et al., 1996; Rossouw et al., 1997), the modified McLaughlin technique and subacromial decompression is still widely
722
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
used for rotator cuff repair (Cordasco and Bigliani, 1997; Hata et al., 2001; Hawkins et al., 1985; Jost et al., 2000; Worland et al., 1999). Clinical experience has shown good to excellent clinical outcomes for rotator cuff repair. Nevertheless, full or partial re-rupture of the repair is one of the main post-operative complications (Jost et al., 2000; Knudsen et al., 1999). The rate of failure is strongly affected by the extension of the tear. For small or medium tears, the clinical outcome shows a re-rupture incidence ranging from 0% to 18% (Bigliani et al., 1992; McKee and Yoo, 2000; Romeo et al., 1999; Cofield et al., 2001; Gazielly et al., 1994; Adamson and Tibone, 1993; Harryman et al., 1991). These values increase in the case of large or massive tears (Cofield et al., 2001; Gazielly et al., 1994; Harryman et al., 1991) with a maximum of 62% (Thomazeau et al., 1997). Although the tear size is the predominant factor, other factors (age, level of activity, gender) may also influence the clinical outcome (Romeo et al., 1999; Gazielly et al., 1994; Harryman et al., 1991; Hattrup, 1995; Wulker et al., 1991). Several authors have investigated the mechanical properties of different rotator cuff repair techniques (France et al., 1989; Gerber et al., 1994; Burkhart et al., 1996; Sward et al., 1992; Caldwell et al., 1997). Experimental methods were used to evaluate the strength of the tendon–bone repair. The results showed that the strength of the repair depends on the repair technique. Three different modes of failure were observed: suture pull-out from the tendon, thread breakage, and sutureto-bone fixation failure. The different modes of failure observed during the in vitro testing show that the tendon-to bone repair has three weak points: the tendon– suture interface, the thread and the suture-to-bone interface. This biomechanical study was performed to assess the current methods of tendon-to-bone repair for rotator cuff tears. In particular, it was meant to investigate if a grasping technique, simpler to be stitched than the Modified Mason–Allen, proposed as standard by other authors (Gerber et al., 1994; Gerber et al., 2000; Gerber et al., 1999), can be used by surgeon in low accessible area of tendon end. Additionally, since the knot is a weak link of the suture loop (Burkhart et al., 1996), an alternative suture knotting technique, that involve fixing of two sutures simultaneously, was proposed and tested to assess if any improvement in the strength to failure of the loop can be achieved. Finally, the effectiveness of the augmentation of the suture-to-bone fixation with plates, proposed by other authors (Gerber et al., 1994; Caldwell et al., 1997), was investigated both in case of osteoporotic bone and good quality cortical bone. Different configurations of the three individual components of the repair (tendon grasping, suture knotting, suture-to-bone fixation) were separately tested. The best combinations in terms of mechanical strength to failure were identified
and tested to compare the mechanical behaviour of the entire repair and to suggest potential improvements in the repair technique.
2. Methods 2.1. Experimental model Shoulders of sheep were selected as the in vitro model for testing of the rotator cuff repair. This model was chosen because of the structural and histological similarity of the sheep infraspinatus tendon to the human supraspinatus tendon (Gerber et al., 1994). The sheep humera and infraspinatus tendons were extracted from 29 mature sheep. Each specimen was wrapped in gauze soaked in a saline solution and immediately frozen ()20 C). The following details were common to all of the individual testing methods, as described in detail below. Prior to testing, the fresh-frozen tendons and bones were thawed in a saline solution maintained at 23 C. Then the tendons were sectioned at the insertion site into the humera to represent a tendon avulsion. All samples (tendons, bones and sutures) were wet before testing but all the tests were performed in air at room temperature using a material testing machine (Mod. 8502, Instron Corp., Canton, MA, USA). An extension rate of 6 mm/ min was chosen to allow instantaneous visualisation of the failure mode. The tensile strength of the investigated system component or of the entire repair was defined as the maximum tensile load or the load value corresponding to the first point at which the slope of the load–displacement curve was zero. The latter criterion was fixed to detect the point at which the suture started to cut through the tissue. A non-absorbable braided polyester suture (Ethibond Excel 2, Johnson & Johnson Intl, Brussels, Belgium) was used in all tests. Power analysis on the fly was used to calculate the sample size required to gain a statistical power of at least 0.90 in detecting a 20% difference in the mechanical behaviour of the single component or of the entire repair (Armitage and Berry, 1994; Becker, 1988), but a minimum of five test repetitions was performed for each test configuration. 2.2. Testing of tendon grasping technique Three tendon grasping techniques (Iannotti and Bigliani, 1998) were examined in this study (Fig. 1): the Modified Southern California Orthopadic Institute (S.C.O.I.) stitch, the Modified Mason–Allen stitch and the Mattress stitch. For each test, one stitch was placed about 10 mm proximal to the end of the tendon. A specially designed clamp, with interdigitating teeth and a clamping area of 50 · 20 mm2 , served as the grasping device for the muscular end. Approximately 20–25 mm
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
723
2.3. Testing of suture knotting
Fig. 1. The three tendon-grasping techniques investigated in the present study.
of tendon protruded from the clamp at the sutured end. When the stitch was made, the two thread ends were knotted with a 2111 knot (hereinafter called 2111 Ôsingle knotÕ) around a pulley fixed to the testing machine so as to produce a preload of about 30 N. This is the typical force necessary for pulling a torn tendon back to its attachment site (Gerber et al., 1994). The elongation between tendon and suture was measured using an extensometer (Mod. 2620-601, Instron Corp., Canton, MA, USA) to assess the gap formation due to elastic behaviour of the tendon and the thread, or slippage of the stitch within the tendon. One arm of the extensometer was fixed 5 mm above the stitch by a needle. The second arm was attached to the suture with a rubber gripping device (Fig. 2). The test was performed following the condition described above, and the tensile strength was measured. Ten infraspinatus tendon specimens were tested for each grasping technique.
Two different techniques of knotting the thread were investigated in this study (Fig. 3). The first involved knotting the two free ends of a single suture using the 2111 knot technique; this is referred to as the 2111 Ôsingle knotÕ. The second method consisted of knotting the four free ends of a pair of sutures using the 2111 knot technique, subsequently referred to as the 2111 Ôdouble knotÕ. To compare the 2111 Ôsingle-knotÕ with the 2111 Ôdouble-knotÕ, a specially designed array of four pulleys was manufactured (Fig. 4). The pulleys had a diameter of 20 mm. The initial distance between the centerpoints of the pulleys was 120 mm. This corresponded to the set-up of Gerber et al. (1994). The 2111 Ôsingle-knotÕ was tested in two ways: (a) only one suture was knotted around two pulleys to form a loop; (b) two threads were tied parallel forming two loops (Fig. 4). As in the previous test, the 2111 knot was tied by a surgeon producing a preload of 30 N. For the two parallel 2111 Ôsingle-knotsÕ, the first knot was tied under a preload of 15 N, and the second knot was then tied to produce a total preload of 30 N. In this way, an equal distribution of the load on the two suture loops was achieved. Since the accuracy of preloading the suture was imprecise, the load was adjusted to the correct value by individually tightening or loosening the screws of the two moving pulleys (Fig. 4). Five test repetitions were performed for each configuration measuring the tensile strength, following the common procedure described above. 2.4. Testing of suture-to-bone fixation A double transosseous suture bone fixation technique was studied. Three different configurations were investigated. The first configuration involved fixation of the suture to the humeral bone of good quality. The second configuration was identical to the previous one, except that the cortical bone around the attachment site was
Fig. 2. The experimental setup used to test the strength of the tendongrasping technique: (A) grasping device for the tendon muscular end; (B) tendon; (C) extensometer; (D) pulley. Preliminary tests were performed to verify the rubber gripping device. No slippage can occur during testing because the minimum force required for moving the rubber on the suture was 1.5 times the maximum force originated by the spring of the extensometer.
Fig. 3. The two suture knotting investigated in the present study: on the left two Ôsingle knotsÕ working in parallel, on the right the Ôdouble knotÕ.
724
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
Fig. 4. The experimental setup used to test the strength of the suture knotting technique: (A) adjusting screws; (B) moving pulleys; (C) fixed pulleys.
modified to simulate osteoporotic bone. This involved removal of the cortical bone at the greater tuberosity until cancellous bone was exposed. The third configuration was as for the second case, except that the attachment was augmented with metallic plate. Two commercially available plates were used: the Button Plate 7 Holes (Synthes, Stratec Medical, Oberdorf, Switzerland), denoted by BP7H, made of pure titanium; and the Plate for Shoulder with 2 holes (Citieffe S.r.l., Bologna, Italy), denoted by PS2H, made of Ti6Al4V. The experimental setup excluded the tendon, solely testing the suture-to-bone attachment. In this way, the strength of the suture-to-bone fixation can be determined. The thawed sheep humera were cleaned of all soft tissue. The distal part of the bone was embedded in a metal box using polymethylmethacrylate bone cement, leaving the humeral head and diaphysis protruding. This constraint left enough space for fixing the suture. The specimen was aligned so that the repaired humerus was oriented with the frontal plane sloped upwards by 15 from the horizontal and the anterior surface superior (Fig. 5). The relative angles between the two suture loops were 90 in the transversal plane and 75 in the sagittal plane. The angle in the frontal plane was not fixed specifically: the bone was mounted on the testing machine such that the lateral surface of the tuberculum minus was parallel to the threads of the suture. Two tunnels were drilled into the tuberosity with a 3 mm diameter drill. Each tunnel was produced by drilling two holes at obtuse angle to meet inside the bone to fuse to a single tunnel. The lateral openings of the two tunnels were positioned directly at the former attachment
Fig. 5. The experimental setup used to test the strength of the sutureto-bone fixation technique: (A) adjusting screws; (B) moving pulleys; (C) extensometer.
site of the infraspinatus tendon. This method was used to produce profound tunnels instead of straight tunnels that run close to the surface of the bone. The spacing distance between the two tunnels (about 10 mm) was fixed following the manufacturer instructions. To perform the suture-to-bone tests the suture was passed over an array of pulleys, through the bone tunnels and tied by a 2111 Ôdouble knotÕ. This type of knot was chosen on the basis of the results of the previous test series, reported below in Section 3. The knot was centred in the spacing distance between the two tunnels and tightened to preload the threads to about 30 N. The load was adjusted to the exact value using the adjusting screws, as before. A monotonic tensile test was performed to measure the tensile strength of the system. Elongation of the suture-to-bone fixation was measured by an extensometer (Mod. 2620–601, Instron Corp.). One arm of the extensometer was equipped with a thin needle that was inserted into a 0.5 mm drill-hole between the two lateral openings of the tunnels. The second arm was attached to the suture with a rubber gripping device. The four configurations were tested five times. 2.5. Testing of entire repair construct Finally, the entire repair construct was tested. These tests were aimed at determining the tensile strength and the Ôweakest linkÕ of the entire repair. Four configurations of the complete system (composed of tendon grasping, suture knotting, suture-to-bone fixation) were investigated. Two Modified S.C.O.I. or two Modified Mason–Allen stitches were used as the tendon grasping technique. For each of these two grasping techniques,
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
725
post-hocs analysis (ScheffeÕs test) were used to analyse the experimental data. The level of significance was set to 0.05. Otherwise, the Kruskal–Wallis test and subsequent non-parametric multiple comparisons (N-P.M.C.) were applied. In this case, the experimentwise error rate was set to obtain a level of significance P ¼ 0:05, as in the previous case.
3. Results 3.1. Testing of tendon grasping technique
Fig. 6. The experimental setup used to test the strength of the entire repair construct: (A) grasping device for the tendon muscular end; (B) extensometer.
the simulated osteoporotic bone was augmented using either the BP7H plate or the PS2H plate. These techniques were selected on the basis of the results of the previous test series. After sewing the two stitches on the tendon, both suture ends of each thread were passed transosseousely and tied over the augmentation plate on the tuberosity using a 2111 Ôdouble-knotÕ (Fig. 6). The knot was tied for a preload of 30 N applied to the tendon. The specially designed clamp (described above) served as the grasping device for the muscular end of the tendon. The humera were oriented as in the previous experiment for testing the suture-to-bone fixation. However, for the angle in the frontal plane the pot was rotated on the machine plate until the flat surface of the tendon and the surface of the tuberosity on which the infraspinatus tendon initially took its origin were parallel. The entire system was evaluated in term of compliance and strength. Elongation of tendon-to-bone fixation was measured by an extensometer (Mod. 2620-601, Instron Corp.). One arm of the extensometer was equipped with a thin needle that was inserted into a 0.5 mm drill-hole between the two lateral openings of the tunnels. The second needle of the extensometer was pinned into the tendon about 5 mm above the stitch, creating an initial gauge-length of about 15 mm. Seven repetitions of the test were performed on each configuration. 2.6. Statistical analysis The statistical analysis applied depended on the type of experimental data. If the different groups were homoschedastic then the analysis of variance (A N O V A ) and
Two different modes of failure were observed for the three investigated tendon grasping techniques. The Modified S.C.O.I. stitch and the Modified Mason–Allen stitch failed by rupture of the suture. In all cases the failure occurred at the level of the 2111 Ôsingle knotÕ. At the end of the test, little or negligible slipping was noted at the insertion points of the suture into the tendon. The Mattress tendon grasping technique failed by cutting through the tendon in 8 out of 10 cases. In the remaining 2 cases, failure was caused by rupture of the suture at the knot level, as observed for the previous stitches. The measured tensile strengths of the three grasping techniques reflect the experimental findings about the failure mode (Fig. 7A). The Kruskal–Wallis test showed significant differences among the three groups (P ¼ 0:007). The tensile strength of the Mattress stitch was lower than the tensile strength of the other two stitches (Mattress vs. Modified Mason–Allen, N-P.M.C., P ¼ 0:003, Mattress vs. Modified S.C.O.I., N-P.M.C., P ¼ 0:017) while no difference was observed between the Modified S.C.O.I. stitch and the Modified Mason–Allen stitch (N-P.M.C., P > 0:05). 3.2. Testing of suture knotting Fig. 7B shows the tensile strength measured for the three investigated configurations. In all cases, failure occurred at the knot level. There are two points worth noting about these results. First, that the failure mode of the 2111 Ôsingle knotÕ was identical to the failure mode of the tendon–suture system observed in the previous test series for the Modified S.C.O.I. grasping technique and the Modified Mason–Allen grasping technique. In fact, the mean value measured for the 2111 Ôsingle knotÕ, the Modified S.C.O.I. grasping technique and the Modified Mason–Allen were 119, 116, and 119 N, respectively. Second, that the tensile strength for the 2111 Ôsingle knotÕ was approximately half the value determined for both the two 2111 Ôsingle knotÕ and the 2111 Ôdouble knotÕ (ScheffeÕs test, P < 0:001 for both comparisons). The difference between the last two mean values was not statistically significant (ScheffeÕs test, P > 0:05).
726
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
Fig. 7. Tensile strength (the mean value of each group is reported) of the single components of the repair and of the entire repair: (A) tensile strength of the tendon-grasping techniques; (B) tensile strength of the suture knotting techniques; (C) tensile strength of the suture-to-bone fixation techniques; (D) tensile strength of the entire repair construct.
3.3. Testing of suture-to-bone fixation The results for the tensile strength of the suture-tobone fixation are summarised in Fig. 7C. Failure of the suture occurred in three of the investigated configurations: the good quality cortical bone (unaugmented), and both the augmented osteoporotic simulated bone conditions. The failure of the suture, tied using a 2111 Ôdouble knotÕ, was at the level of the knot. Conversely, the unaugmented osteoporotic simulated bone fixation failed by yielding the bone, where the suture did not break, cutting through the bone. The statistical analysis showed that there was a significant difference (A N O V A , P < 0:001) between the four groups. No statistical difference was found between the following three fixations: suture-unaugmented good quality cortical bone, sutureBP7H augmented osteoporotic simulated bone, and suture-PS2H augmented osteoporotic simulated bone. All three fixations were significantly more resistant than the suture-unaugmented osteoporotic simulated bone system (ScheffeÕs test, P < 0:001 for all the three comparisons). 3.4. Testing of entire repair construct The initial response of the repair to loading was the same for all four investigated configurations. As the distracting displacement was applied to the tendon, the tendon and the suture were stretched, while the plate
was compressed against the bone. Initially, with an increase in the applied displacement, the load increase was almost linear. The slope of the linear portion of the elongation–force curve was assumed to represent the compliance of the tendon–suture–bone system (Table 1). The statistical analysis showed no difference between the mean compliance values of the four groups (A N O V A , P > 0:05). A further increase in the distracting displacement produced a gap larger than 10 mm between the tendon and the bone, which was observed before reaching the maximum tensile strength in all investigated repairs. Failure of the repair always occurred at the stitch–tendon interface. Two failure modes were observed: rupture of the suture or pull out of the suture from the tendon. Of the 14 repairs performed using the Modified S.C.O.I. grasping technique, 6 failed by failure of the thread at the stitch level and the other 8 failed by suture pull out of the tendon. All the 14 repairs performed using the Modified S.C.O.I. grasping technique failed by suture pull out of the tendon. Fig. 7D shows the tensile strength measured for the four investigated configurations. The plate used for bone augmentation had no statistical effect (A N O V A , P > 0:05) on the tensile strength of the repair. In addition, no statistical difference (A N O V A , P > 0:05) was found between the mean values of the tensile strength calculated for the two tendon grasping techniques, in spite of the different modes of failure observed.
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
727
Table 1 Compliance of the entire repair construct calculated for the linear part of the load–displacement curve Configuration Bone quality Simulated Simulated Simulated Simulated
osteoporotic osteoporotic osteoporotic osteoporotic
bone bone bone bone
Grasping technique
Button plate type
Two Two Two Two
BP7H BP7H PS2H PS2H
Mod. Mod. Mod. Mod.
Mason–Allen S.C.O.I. Mason–Allen S.C.O.I.
4. Discussion The present study investigates the mechanical properties of different techniques for rotator cuff repair. The proposed experimental method used an animal model, testing the infraspinatus tendon and humerus of sheep. The sheep model has already been used by others to investigate the mechanical strength of different rotator cuff repair techniques. In fact, the histological and mechanical characteristics of sheep infraspinatus tendon are similar to those of human supraspinatus tendons (Gerber et al., 1994; Lewis et al., 2001). It is recognised that there are a number of limitations in this study. The experimental method cannot simulate degeneration of the tendon that may be observed in delayed repair in human. All the samples (tendon-humerus) were collected intact and, at the time of testing, the tendon was transected to represent an acute tear. While this does not represent an actual tear, the method of transection was preferred, being repeatable for every specimen. In addition, the sheep humera were not osteoporotic. The cortical bone of the greater tuberosity was removed to simulate the effect of osteoporosis. Again, this procedure was used to ensure repeatability. Finally, the experimental setup did not simulate the physiological activities that can stress the repair in vivo. Therefore, the absolute values of the repair tensile strength found in this study cannot be directly transferred to clinically relevant numbers. Nevertheless, the experimental data collected can be useful to highlight the difference between various rotator cuff repair techniques. One purpose of the present study was to compare the holding power of three grasping techniques: the Modified S.C.O.I., the Modified Mason–Allen, and the Mattress. The failure mode of the tendon–suture system was identified as suture failure or tendon cut through. The experimental data showed that the Mattress group had a lower load to failure than the remaining two groups. Out of the 10 tendon–suture constructs of the Mattress group, the suture cut through the tendon in 8. Conversely, the Modified S.C.O.I. and the Modified Mason–Allen did not slip within the tendon. In all cases, failure occurred for suture breakage at the thread knot. Hence, the holding power of the Modified Mason–Allen stitch and of the Modified S.C.O.I. stitch is larger than that measured for the Mattress stitch.
Entire repair compliance (mm/N) Mean (S.D.) 0.046 0.048 0.045 0.047
(0.013) (0.009) (0.014) (0.011)
The recurrence of tendon–suture construct failure at the thread knot level suggests that the knot is a weak link of the suture loop. Gerber et al. (1994) reported an ultimate tensile strength of 82 ± 3 N for a single Ethibond no. 2 thread. In the present study the suture loop tied together by a single knot failed at about 120 N. Since the load was equally distributed on the two threads of the suture loop (the load was equally distributed over the pulley), it can be calculated that breakage of the suture occurred at about 60 N. This reduced value of the tensile strength confirms that the knot weakened the thread, as already suggested by Burkhart et al. (1996). Two different reasons explain this observation: the knot caused a stress concentration in the thread, the thread was damaged at the knot level during tying. To improve this weak link of the suture loop, the Ôdouble knotÕ was introduced. This solution showed a tensile strength over twice that of the Ôsingle knotÕ (252 and 119 N, respectively) but the Ôdouble knotÕ involve fixing two sutures. No significant difference was found in the tensile strength values of the Ôdouble knotÕ and of two Ôsingle knotsÕ working in parallel. Hence, the Ôdouble knotÕ does not improve the tensile strength of the suture locking. Additionally, the Ôdouble knotÕ may require a greater effort by the surgeon to reduce the gap between the tendon and the bone surface. In fact, knotting two stitches simultaneously implies that a wider portion of the contracted tendon needs to be pulled to bone than that in case of a Ôsingle knotÕ. Nevertheless, the double knot seems preferable because this solution would reduce differences in the load stressing the two suture loops tied together and finally should distribute the load on the tendon more homogeneously. The results of the test series on suture-to-bone fixation showed that fixation over augmented bone was superior than that over unaugmented bone when the cortical bone has poor mechanical characteristics (osteoporotic bone). These findings are in agreement with previous reports (Sward et al., 1992; Caldwell et al., 1997). The metal plate increases the contact area between the repair and the bone reducing the compressive pressure on the bone and finally the risk of bone failure. Conversely, the bone augmentation seems mechanically useless when fixation is performed on good quality cortical bone. In fact, in the present investigation good quality cortical bone withstood compressive pressure
728
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
until suture failure occurred, with negligible tissue damage at the contact area. It is surprising that in this test series the Ôdouble knotÕ failed at a higher load than that measured in testing the knot suture alone. Theoretically, the strength of the repair without tendon should not be higher than that of the Ôdouble knotÕ. This observation suggests the hypothesis that the load is partially transmitted from the suture to the bone within the tunnels by friction. Therefore, the knot (the weak point of the suture loop), being located on the far side of the tunnels, is only stressed by a fraction of the load stressing the threads. Thus, higher loads can be applied to the suture before reaching the ultimate tensile strength of the knot. For all of the investigated entire repair constructs, the bone–suture fixation involved a Ôdouble knotÕ tied over a plate augmentation. The entire repairs never failed over the metallic plate or at the suture knot level. Hence, the bone augmented fixation technique transferred the weakest point of the repair to the suture–tendon interface or at the suture material. These experimental results are in agreement with previous findings (Gerber et al., 1994; Sward et al., 1992). The elongation under tensile load was not different for the investigated repair techniques. The large compliance of the four entire repairs was unexpected. A large gap between the tendon end and the bone surface of up to 10 mm was observed prior to failure of the repair. Hence, physiological loads may generate a gap at the tendon–bone interface, where values of up to 300 N have been reported in the supraspinatus tendon for abduction of the arm (Wallace, 1984). The findings of Gerber et al. (1999) and Burkhart et al. (1997a,b) support this statement. In testing the entire repair, the two investigated grasping techniques, the Modified Mason–Allen stitch and the Modified S.C.O.I., showed a comparable tensile strength. However, two failure modes of the tendon– suture system were identified: rupture of the suture at the stitch level or pull out of the suture from the tendon. In all cases, the Modified S.C.O.I. repair failed by cutting through the tendon. Of the 14 Modified Mason– Allen repairs tested, 8 failed by cutting through the tendon, and the other 6 failed by suture breakage at the stitch level. These findings would suggest that the Modified Mason–Allen stitch has a higher tendon holding power than the Modified S.C.O.I. stitch, but at the same time this results in a higher stress level in the suture material for the Modified Mason–Allen stitch in comparison with stress level for the Modified S.C.O.I. stitch. Both the holding power and the stress level in the suture material are an intrinsic characteristic of the stitch affected by the tendon tissue quality and by thread size. The surgeon cannot change the tissue quality of the patient tendon. Conversely, the suture size can be increased with no need for larger transosseous tunnels. Considering the Modified Mason–Allen stitch strength
from this study and the results of Gerber et al. (1994), it can be seen that an increase in one size of the suture can increase the tensile strength by 50%. The compliance of the repair would also decrease. Hence, it would be recommended to use a suture size larger than no. 2. However, a limit exists in increasing the size of thread: too larger a suture may promote local irritation due to possible contact with the acromion. In addition, a tendon–suture augmentation would improve the tensile strength of the repair (France et al., 1989). Nevertheless, adding an augmentation patch at the tendon–thread interface may cause impingement (Sward et al., 1992).
5. Conclusions The use of ÔreinforcedÕ stitches (Modified Mason– Allen stitch or Modified S.C.O.I. stitch) for tendon grasping seems recommendable due to the higher tensile strength of these stitches in comparison with the Mattress stitch. The Ôdouble knotÕ locking, reducing a possible unequal load distribution between two suture loops, seems preferable to two Ôsingle knotsÕ working in parallel. The suture-to-bone fixation over augmented cortical bone had a higher mechanical strength than fixation over osteoporotic bone. Conversely, augmentation over good quality cortical bone seemed to produce a negligible improvement of the suture-to-bone fixation strength. Nevertheless of the efforts to improve the surgical technique and finally the strength of the repair, the compliance of the repair remain a problem. Post-operative protection of the repaired rotator cuff from tension load is necessary.
Acknowledgements The authors wish to thank Luigi Lena for the illustrations, Roberta Fognani for technical support during the experiments, Cristina Ancarani for the statistical analysis and Victor Waide for help with the English revision.
References Adamson, G.J., Tibone, J.E., 1993. Ten-year assessment of primary rotator cuff repairs. J. Shoulder Elbow Surg. 2, 57–63. Armitage, P., Berry, G., 1994. Statistical Methods in Medical Research, third ed. Blackwell Scientific Publications, Oxford. Becker, B.J., 1988. Synthesizing standardized mean-change measures. Brit. J. Math. Stat. Psychol. 41, 257–278. Bigliani, L.U., Kimmel, J., McCann, P.D., Wolfe, I., 1992. Repair of rotator cuff tears in tennis players. Am. J. Sports Med. 20, 112–117. Burkhart, S.S., Fischer, S.P., Nottage, W.M., Esch, J.C., Barber, F.A., Doctor, D., et al., 1996. Tissue fixation security in transosseous
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729 rotator cuff repairs: a mechanical comparison of simple versus mattress sutures. Arthroscopy 12, 704–708. Burkhart, S.S., Diaz Pagan, J.L., Wirth, M.A., Athanasiou, K.A., 1997a. Cyclic loading of anchor-based rotator cuff repairs: confirmation of the tension overload phenomenon and comparison of suture anchor fixation with transosseous fixation. Arthroscopy 13, 720–724. Burkhart, S.S., Johnson, T.C., Wirth, M.A., Athanasiou, K.A., 1997b. Cyclic loading of transosseous rotator cuff repairs: tension overload as a possible cause of failure. Arthroscopy 13, 172–176. Caldwell, G.L., Warner, J.P., Miller, M.D., Boardman, D., Towers, J., Debski, R., 1997. Strength of fixation with transosseous sutures in rotator cuff repair. J. Bone Joint Surg Am. 79, 1064–1068. Cofield, R.H., Parvizi, J., Hoffmeyer, P.J., Lanzer, W.L., Ilstrup, D.M., Rowland, C.M., 2001. Surgical repair of chronic rotator cuff tears. A prospective long-term study. J. Bone Joint Surg Am. 83-A, 71–77. Cordasco, F.A., Bigliani, L.U., 1997. The rotator cuff. Large and massive tears. Technique of open repair. Orthop. Clin. North Am. 28, 179–193. Craft, D.V., Moseley, J.B., Cawley, P.W., Noble, P.C., 1996. Fixation strength of rotator cuff repairs with suture anchors and the transosseous suture technique. J. Shoulder Elbow Surg. 5, 32–40. France, E.P., Paulos, L.E., Harner, C.D., Straight, C.B., 1989. Biomechanical evaluation of rotator cuff fixation methods. Am. J. Sports Med. 17, 176–181. Gazielly, D.F., Gleyze, P., Montagnon, C., 1994. Functional and anatomical results after rotator cuff repair. Clin. Orthop. 304, 43– 53. Gerber, C., Schneeberger, A.G., Beck, M., Schlegel, U., 1994. Mechanical strength of repairs of the rotator cuff. J. Bone Joint Surg. Br. 76, 371–380. Gerber, C., Schneeberger, A.G., Perren, S.M., Nyffeler, R.W., 1999. Experimental rotator cuff repair. A preliminary study. J. Bone Joint Surg Am. 81, 1281–1290. Gerber, C., Fuchs, B., Hodler, J., 2000. The results of repair of massive tears of the rotator cuff. J. Bone Joint Surg Am. 82, 505–515. Goradia, V.K., Mullen, D.J., Boucher, H.R., Parks, B.G., OÕDonnell, J.B., 2001. Cyclic loading of rotator cuff repairs: a comparison of bioabsorbable tacks with metal suture anchors and transosseous sutures. Arthroscopy 17, 360–364. Harryman, D.T., Mack, L.A., Wang, K.Y., Jackins, S.E., Richardson, M.L., Matsen, F.A., 1991. Repairs of the rotator cuff. Correlation of functional results with integrity of the cuff. J. Bone Joint Surg Am. 73, 982–989. Hata, Y., Saitoh, S., Murakami, N., Seki, H., Nakatsuchi, Y., Takaoka, K., 2001. A less invasive surgery for rotator cuff tear: mini-open repair. J. Shoulder Elbow Surg. 10, 11–16. Hattrup, S.J., 1995. Rotator cuff repair: relevance of patient age. J. Shoulder Elbow Surg. 4, 95–100.
729
Hawkins, R.J., Misamore, G.W., Hobeika, P.E., 1985. Surgery for full-thickness rotator-cuff tears. J. Bone Joint Surg. Am. 67, 1349– 1355. Hecker, A.T., Shea, M., Hayhurst, J.O., Myers, E.R., Meeks, L.W., Hayes, W.C., 1993. Pull-out strength of suture anchors for rotator cuff and Bankart lesion repairs. Am. J. Sports Med. 21, 874–879. Iannotti, J.P., Bigliani, L.U., 1998. The Rotator Cuff: Current Concepts and Complex Problems. American Academy of Orthopaedic Surgeon, Rosemont. Jost, B., Pfirrmann, C.W., Gerber, C., Switzerland, Z., 2000. Clinical outcome after structural failure of rotator cuff repairs. J. Bone Joint Surg Am. 82, 304–314. Knudsen, H.B., Gelineck, J., Sojbjerg, J.O., Olsen, B.S., Johannsen, H.V., Sneppen, O., 1999. Functional and magnetic resonance imaging evaluation after single-tendon rotator cuff reconstruction. J. Shoulder Elbow Surg. 8, 242–246. Lewis, C.W., Schlegel, T.F., Hawkins, R.J., James, S.P., Turner, A.S., 2001. The effect of immobilization on rotator cuff healing using modified Mason–Allen stitches: a biomechanical study in sheep. Biomed. Sci. Instrum. 37, 263–268. McKee, M.D., Yoo, D.J., 2000. The effect of surgery for rotator cuff disease on general health status. Results of a prospective trial. J. Bone Joint Surg. Am. 82-A, 970–979. McLaughlin, H.L., 1944. Lesion of the musculotendinous cuff of the shoulder. The exposure and treatment of tears with retraction. J. Bone Joint Surg. Am. 26, 31–51. Reed, S.C., Glossop, N., Ogilvie-Harris, D.J., 1996. Full-thickness rotator cuff tears. A biomechanical comparison of suture versus bone anchor techniques. Am. J. Sports Med. 24, 46–48. Romeo, A.A., Hang, D.W., Bach, B.R., Shott, S., 1999. Repair of full thickness rotator cuff tears. Gender, age, and other factors affecting outcome. Clin. Orthop. 367, 243–255. Rossouw, D.J., McElroy, B.J., Amis, A.A., Emery, R.J., 1997. A biomechanical evaluation of suture anchors in repair of the rotator cuff. J. Bone Joint Surg. Br. 79, 458–461. Sward, L., Hughes, J.S., Amis, A., Wallace, W.A., 1992. The strength of surgical repairs of the rotator cuff. A biomechanical study on cadavers. J. Bone Joint Surg. Br. 74, 585–588. Thomazeau, H., Boukobza, E., Morcet, N., Chaperon, J., Langlais, F., 1997. Prediction of rotator cuff repair results by magnetic resonance imaging. Clin. Orthop. 344, 275–283. Wallace, W.A., 1984. Evaluation of the force, the I.C.R. and the neutral point during abduction of the shoulder. In: 30th Annual Meeting of the Orthopaedic Research Society, Atlanta, Georgia. Worland, R.L., Arredondo, J., Angles, F., Lopez_Jimenez, F., 1999. Repair of massive rotator cuff tears in patients older than 70 years. J. Shoulder Elbow Surg. 8, 26–30. Wulker, N., Melzer, C., Wirth, C.J., 1991. Shoulder surgery for rotator cuff tears. Ultrasonographic 3-year follow-up of 97 cases. Acta Orthop. Scand. 62, 142–147.
Surgical repair of the rotator cuff: a biomechanical evaluation of different tendon grasping and bone suture fixation techniques Massimiliano Baleani a
a,*
, Stephan Schrader a, Carlo Andrea Veronesi b, Roberto Rotini b, Roberto Giardino c, Aldo Toni a,d
Laboratorio di Tecnologia Medica, Istituti Ortopedici Rizzoli, Via Di Barbiano 1/10, 40136 Bologna, Italy b Modulo Dipartimentale Chirurgia Spalla-Gomito, Istituti Ortopedici Rizzoli, 40136 Bologna, Italy c Servizio di Chirurgia Sperimentale, Istituti Ortopedici Rizzoli, 40136 Bologna, Italy d I Divisione Ortopedia e Traumatologia, Istituti Ortopedici Rizzoli, 40136 Bologna, Italy Received 28 January 2002; accepted 28 May 2003
Abstract Objective. This study investigated the initial strength and failure mode of different rotator cuff repair techniques. Background. Full or partial re-rupture of the repair is one of the main post-operative complications for rotator cuff repair. The rate of failure is strongly affected by the extension of the tear, increasing in case of large or massive tears up to 62%. Design. The study was planned to assess the three individual components of the tendon-to-bone repair (tendon grasping, suture knotting, suture-to-bone fixation) and to identify the best combinations in terms of mechanical strength to failure. The best combinations were tested to compare the mechanical behaviour of the entire repair and suggest potential improvements in the repair technique. Methods. Experimental tests were performed using sheep shoulders. Three tendon-grasping techniques, two suture knotting techniques, and the effect of bone augmentation with metallic plate and bone quality on suture-to-bone fixation were investigated. Results. This study assessed the mechanical behaviour of different repair components. The best combinations of the investigated techniques showed that the weakest link was the tendon–suture interface. More importantly, the compliance of the investigated repairs was large. Conclusions. The initial strength of the rotator cuff repair can be improved by changing the repair technique. Nevertheless, even a low physiological load stressing the repaired tendon may cause a gap formation at the tendon–bone interface without necessarily producing failure of the repair. Relevance Post-operative protection of the repaired rotator cuff from tension load is necessary to reduce the risk of delaying or preventing of the healing process. 2003 Elsevier Ltd. All rights reserved. Keywords: Rotator cuff; Repair technique; Biomechanics; Experimental testing; Tensile strength; Stiffness
1. Introduction Lesion of the rotator cuff is a common problem in orthopaedics. Progressive pain or functional deficits to the shoulder joint generally requires surgical repair of the rotator cuff. Traditional techniques for the repair of the rotator cuff have evolved from the work of
*
Corresponding author. E-mail address: [email protected] (M. Baleani).
0268-0033/$ - see front matter 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0268-0033(03)00122-0
McLaughlin (1944). McLaughlin described a surgical method for the reattachment of the avulsed rotator cuff tendon to the humeral bone by passing sutures through transosseous tunnels in the greater tuberosity. Although others have subsequently studied the use of different fixation techniques, trying to obtain a greater strength of fixation or to define a less invasive technique (Craft et al., 1996; France et al., 1989; Goradia et al., 2001; Hecker et al., 1993; Gerber et al., 1994; Reed et al., 1996; Rossouw et al., 1997), the modified McLaughlin technique and subacromial decompression is still widely
722
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
used for rotator cuff repair (Cordasco and Bigliani, 1997; Hata et al., 2001; Hawkins et al., 1985; Jost et al., 2000; Worland et al., 1999). Clinical experience has shown good to excellent clinical outcomes for rotator cuff repair. Nevertheless, full or partial re-rupture of the repair is one of the main post-operative complications (Jost et al., 2000; Knudsen et al., 1999). The rate of failure is strongly affected by the extension of the tear. For small or medium tears, the clinical outcome shows a re-rupture incidence ranging from 0% to 18% (Bigliani et al., 1992; McKee and Yoo, 2000; Romeo et al., 1999; Cofield et al., 2001; Gazielly et al., 1994; Adamson and Tibone, 1993; Harryman et al., 1991). These values increase in the case of large or massive tears (Cofield et al., 2001; Gazielly et al., 1994; Harryman et al., 1991) with a maximum of 62% (Thomazeau et al., 1997). Although the tear size is the predominant factor, other factors (age, level of activity, gender) may also influence the clinical outcome (Romeo et al., 1999; Gazielly et al., 1994; Harryman et al., 1991; Hattrup, 1995; Wulker et al., 1991). Several authors have investigated the mechanical properties of different rotator cuff repair techniques (France et al., 1989; Gerber et al., 1994; Burkhart et al., 1996; Sward et al., 1992; Caldwell et al., 1997). Experimental methods were used to evaluate the strength of the tendon–bone repair. The results showed that the strength of the repair depends on the repair technique. Three different modes of failure were observed: suture pull-out from the tendon, thread breakage, and sutureto-bone fixation failure. The different modes of failure observed during the in vitro testing show that the tendon-to bone repair has three weak points: the tendon– suture interface, the thread and the suture-to-bone interface. This biomechanical study was performed to assess the current methods of tendon-to-bone repair for rotator cuff tears. In particular, it was meant to investigate if a grasping technique, simpler to be stitched than the Modified Mason–Allen, proposed as standard by other authors (Gerber et al., 1994; Gerber et al., 2000; Gerber et al., 1999), can be used by surgeon in low accessible area of tendon end. Additionally, since the knot is a weak link of the suture loop (Burkhart et al., 1996), an alternative suture knotting technique, that involve fixing of two sutures simultaneously, was proposed and tested to assess if any improvement in the strength to failure of the loop can be achieved. Finally, the effectiveness of the augmentation of the suture-to-bone fixation with plates, proposed by other authors (Gerber et al., 1994; Caldwell et al., 1997), was investigated both in case of osteoporotic bone and good quality cortical bone. Different configurations of the three individual components of the repair (tendon grasping, suture knotting, suture-to-bone fixation) were separately tested. The best combinations in terms of mechanical strength to failure were identified
and tested to compare the mechanical behaviour of the entire repair and to suggest potential improvements in the repair technique.
2. Methods 2.1. Experimental model Shoulders of sheep were selected as the in vitro model for testing of the rotator cuff repair. This model was chosen because of the structural and histological similarity of the sheep infraspinatus tendon to the human supraspinatus tendon (Gerber et al., 1994). The sheep humera and infraspinatus tendons were extracted from 29 mature sheep. Each specimen was wrapped in gauze soaked in a saline solution and immediately frozen ()20 C). The following details were common to all of the individual testing methods, as described in detail below. Prior to testing, the fresh-frozen tendons and bones were thawed in a saline solution maintained at 23 C. Then the tendons were sectioned at the insertion site into the humera to represent a tendon avulsion. All samples (tendons, bones and sutures) were wet before testing but all the tests were performed in air at room temperature using a material testing machine (Mod. 8502, Instron Corp., Canton, MA, USA). An extension rate of 6 mm/ min was chosen to allow instantaneous visualisation of the failure mode. The tensile strength of the investigated system component or of the entire repair was defined as the maximum tensile load or the load value corresponding to the first point at which the slope of the load–displacement curve was zero. The latter criterion was fixed to detect the point at which the suture started to cut through the tissue. A non-absorbable braided polyester suture (Ethibond Excel 2, Johnson & Johnson Intl, Brussels, Belgium) was used in all tests. Power analysis on the fly was used to calculate the sample size required to gain a statistical power of at least 0.90 in detecting a 20% difference in the mechanical behaviour of the single component or of the entire repair (Armitage and Berry, 1994; Becker, 1988), but a minimum of five test repetitions was performed for each test configuration. 2.2. Testing of tendon grasping technique Three tendon grasping techniques (Iannotti and Bigliani, 1998) were examined in this study (Fig. 1): the Modified Southern California Orthopadic Institute (S.C.O.I.) stitch, the Modified Mason–Allen stitch and the Mattress stitch. For each test, one stitch was placed about 10 mm proximal to the end of the tendon. A specially designed clamp, with interdigitating teeth and a clamping area of 50 · 20 mm2 , served as the grasping device for the muscular end. Approximately 20–25 mm
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
723
2.3. Testing of suture knotting
Fig. 1. The three tendon-grasping techniques investigated in the present study.
of tendon protruded from the clamp at the sutured end. When the stitch was made, the two thread ends were knotted with a 2111 knot (hereinafter called 2111 Ôsingle knotÕ) around a pulley fixed to the testing machine so as to produce a preload of about 30 N. This is the typical force necessary for pulling a torn tendon back to its attachment site (Gerber et al., 1994). The elongation between tendon and suture was measured using an extensometer (Mod. 2620-601, Instron Corp., Canton, MA, USA) to assess the gap formation due to elastic behaviour of the tendon and the thread, or slippage of the stitch within the tendon. One arm of the extensometer was fixed 5 mm above the stitch by a needle. The second arm was attached to the suture with a rubber gripping device (Fig. 2). The test was performed following the condition described above, and the tensile strength was measured. Ten infraspinatus tendon specimens were tested for each grasping technique.
Two different techniques of knotting the thread were investigated in this study (Fig. 3). The first involved knotting the two free ends of a single suture using the 2111 knot technique; this is referred to as the 2111 Ôsingle knotÕ. The second method consisted of knotting the four free ends of a pair of sutures using the 2111 knot technique, subsequently referred to as the 2111 Ôdouble knotÕ. To compare the 2111 Ôsingle-knotÕ with the 2111 Ôdouble-knotÕ, a specially designed array of four pulleys was manufactured (Fig. 4). The pulleys had a diameter of 20 mm. The initial distance between the centerpoints of the pulleys was 120 mm. This corresponded to the set-up of Gerber et al. (1994). The 2111 Ôsingle-knotÕ was tested in two ways: (a) only one suture was knotted around two pulleys to form a loop; (b) two threads were tied parallel forming two loops (Fig. 4). As in the previous test, the 2111 knot was tied by a surgeon producing a preload of 30 N. For the two parallel 2111 Ôsingle-knotsÕ, the first knot was tied under a preload of 15 N, and the second knot was then tied to produce a total preload of 30 N. In this way, an equal distribution of the load on the two suture loops was achieved. Since the accuracy of preloading the suture was imprecise, the load was adjusted to the correct value by individually tightening or loosening the screws of the two moving pulleys (Fig. 4). Five test repetitions were performed for each configuration measuring the tensile strength, following the common procedure described above. 2.4. Testing of suture-to-bone fixation A double transosseous suture bone fixation technique was studied. Three different configurations were investigated. The first configuration involved fixation of the suture to the humeral bone of good quality. The second configuration was identical to the previous one, except that the cortical bone around the attachment site was
Fig. 2. The experimental setup used to test the strength of the tendongrasping technique: (A) grasping device for the tendon muscular end; (B) tendon; (C) extensometer; (D) pulley. Preliminary tests were performed to verify the rubber gripping device. No slippage can occur during testing because the minimum force required for moving the rubber on the suture was 1.5 times the maximum force originated by the spring of the extensometer.
Fig. 3. The two suture knotting investigated in the present study: on the left two Ôsingle knotsÕ working in parallel, on the right the Ôdouble knotÕ.
724
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
Fig. 4. The experimental setup used to test the strength of the suture knotting technique: (A) adjusting screws; (B) moving pulleys; (C) fixed pulleys.
modified to simulate osteoporotic bone. This involved removal of the cortical bone at the greater tuberosity until cancellous bone was exposed. The third configuration was as for the second case, except that the attachment was augmented with metallic plate. Two commercially available plates were used: the Button Plate 7 Holes (Synthes, Stratec Medical, Oberdorf, Switzerland), denoted by BP7H, made of pure titanium; and the Plate for Shoulder with 2 holes (Citieffe S.r.l., Bologna, Italy), denoted by PS2H, made of Ti6Al4V. The experimental setup excluded the tendon, solely testing the suture-to-bone attachment. In this way, the strength of the suture-to-bone fixation can be determined. The thawed sheep humera were cleaned of all soft tissue. The distal part of the bone was embedded in a metal box using polymethylmethacrylate bone cement, leaving the humeral head and diaphysis protruding. This constraint left enough space for fixing the suture. The specimen was aligned so that the repaired humerus was oriented with the frontal plane sloped upwards by 15 from the horizontal and the anterior surface superior (Fig. 5). The relative angles between the two suture loops were 90 in the transversal plane and 75 in the sagittal plane. The angle in the frontal plane was not fixed specifically: the bone was mounted on the testing machine such that the lateral surface of the tuberculum minus was parallel to the threads of the suture. Two tunnels were drilled into the tuberosity with a 3 mm diameter drill. Each tunnel was produced by drilling two holes at obtuse angle to meet inside the bone to fuse to a single tunnel. The lateral openings of the two tunnels were positioned directly at the former attachment
Fig. 5. The experimental setup used to test the strength of the sutureto-bone fixation technique: (A) adjusting screws; (B) moving pulleys; (C) extensometer.
site of the infraspinatus tendon. This method was used to produce profound tunnels instead of straight tunnels that run close to the surface of the bone. The spacing distance between the two tunnels (about 10 mm) was fixed following the manufacturer instructions. To perform the suture-to-bone tests the suture was passed over an array of pulleys, through the bone tunnels and tied by a 2111 Ôdouble knotÕ. This type of knot was chosen on the basis of the results of the previous test series, reported below in Section 3. The knot was centred in the spacing distance between the two tunnels and tightened to preload the threads to about 30 N. The load was adjusted to the exact value using the adjusting screws, as before. A monotonic tensile test was performed to measure the tensile strength of the system. Elongation of the suture-to-bone fixation was measured by an extensometer (Mod. 2620–601, Instron Corp.). One arm of the extensometer was equipped with a thin needle that was inserted into a 0.5 mm drill-hole between the two lateral openings of the tunnels. The second arm was attached to the suture with a rubber gripping device. The four configurations were tested five times. 2.5. Testing of entire repair construct Finally, the entire repair construct was tested. These tests were aimed at determining the tensile strength and the Ôweakest linkÕ of the entire repair. Four configurations of the complete system (composed of tendon grasping, suture knotting, suture-to-bone fixation) were investigated. Two Modified S.C.O.I. or two Modified Mason–Allen stitches were used as the tendon grasping technique. For each of these two grasping techniques,
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
725
post-hocs analysis (ScheffeÕs test) were used to analyse the experimental data. The level of significance was set to 0.05. Otherwise, the Kruskal–Wallis test and subsequent non-parametric multiple comparisons (N-P.M.C.) were applied. In this case, the experimentwise error rate was set to obtain a level of significance P ¼ 0:05, as in the previous case.
3. Results 3.1. Testing of tendon grasping technique
Fig. 6. The experimental setup used to test the strength of the entire repair construct: (A) grasping device for the tendon muscular end; (B) extensometer.
the simulated osteoporotic bone was augmented using either the BP7H plate or the PS2H plate. These techniques were selected on the basis of the results of the previous test series. After sewing the two stitches on the tendon, both suture ends of each thread were passed transosseousely and tied over the augmentation plate on the tuberosity using a 2111 Ôdouble-knotÕ (Fig. 6). The knot was tied for a preload of 30 N applied to the tendon. The specially designed clamp (described above) served as the grasping device for the muscular end of the tendon. The humera were oriented as in the previous experiment for testing the suture-to-bone fixation. However, for the angle in the frontal plane the pot was rotated on the machine plate until the flat surface of the tendon and the surface of the tuberosity on which the infraspinatus tendon initially took its origin were parallel. The entire system was evaluated in term of compliance and strength. Elongation of tendon-to-bone fixation was measured by an extensometer (Mod. 2620-601, Instron Corp.). One arm of the extensometer was equipped with a thin needle that was inserted into a 0.5 mm drill-hole between the two lateral openings of the tunnels. The second needle of the extensometer was pinned into the tendon about 5 mm above the stitch, creating an initial gauge-length of about 15 mm. Seven repetitions of the test were performed on each configuration. 2.6. Statistical analysis The statistical analysis applied depended on the type of experimental data. If the different groups were homoschedastic then the analysis of variance (A N O V A ) and
Two different modes of failure were observed for the three investigated tendon grasping techniques. The Modified S.C.O.I. stitch and the Modified Mason–Allen stitch failed by rupture of the suture. In all cases the failure occurred at the level of the 2111 Ôsingle knotÕ. At the end of the test, little or negligible slipping was noted at the insertion points of the suture into the tendon. The Mattress tendon grasping technique failed by cutting through the tendon in 8 out of 10 cases. In the remaining 2 cases, failure was caused by rupture of the suture at the knot level, as observed for the previous stitches. The measured tensile strengths of the three grasping techniques reflect the experimental findings about the failure mode (Fig. 7A). The Kruskal–Wallis test showed significant differences among the three groups (P ¼ 0:007). The tensile strength of the Mattress stitch was lower than the tensile strength of the other two stitches (Mattress vs. Modified Mason–Allen, N-P.M.C., P ¼ 0:003, Mattress vs. Modified S.C.O.I., N-P.M.C., P ¼ 0:017) while no difference was observed between the Modified S.C.O.I. stitch and the Modified Mason–Allen stitch (N-P.M.C., P > 0:05). 3.2. Testing of suture knotting Fig. 7B shows the tensile strength measured for the three investigated configurations. In all cases, failure occurred at the knot level. There are two points worth noting about these results. First, that the failure mode of the 2111 Ôsingle knotÕ was identical to the failure mode of the tendon–suture system observed in the previous test series for the Modified S.C.O.I. grasping technique and the Modified Mason–Allen grasping technique. In fact, the mean value measured for the 2111 Ôsingle knotÕ, the Modified S.C.O.I. grasping technique and the Modified Mason–Allen were 119, 116, and 119 N, respectively. Second, that the tensile strength for the 2111 Ôsingle knotÕ was approximately half the value determined for both the two 2111 Ôsingle knotÕ and the 2111 Ôdouble knotÕ (ScheffeÕs test, P < 0:001 for both comparisons). The difference between the last two mean values was not statistically significant (ScheffeÕs test, P > 0:05).
726
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
Fig. 7. Tensile strength (the mean value of each group is reported) of the single components of the repair and of the entire repair: (A) tensile strength of the tendon-grasping techniques; (B) tensile strength of the suture knotting techniques; (C) tensile strength of the suture-to-bone fixation techniques; (D) tensile strength of the entire repair construct.
3.3. Testing of suture-to-bone fixation The results for the tensile strength of the suture-tobone fixation are summarised in Fig. 7C. Failure of the suture occurred in three of the investigated configurations: the good quality cortical bone (unaugmented), and both the augmented osteoporotic simulated bone conditions. The failure of the suture, tied using a 2111 Ôdouble knotÕ, was at the level of the knot. Conversely, the unaugmented osteoporotic simulated bone fixation failed by yielding the bone, where the suture did not break, cutting through the bone. The statistical analysis showed that there was a significant difference (A N O V A , P < 0:001) between the four groups. No statistical difference was found between the following three fixations: suture-unaugmented good quality cortical bone, sutureBP7H augmented osteoporotic simulated bone, and suture-PS2H augmented osteoporotic simulated bone. All three fixations were significantly more resistant than the suture-unaugmented osteoporotic simulated bone system (ScheffeÕs test, P < 0:001 for all the three comparisons). 3.4. Testing of entire repair construct The initial response of the repair to loading was the same for all four investigated configurations. As the distracting displacement was applied to the tendon, the tendon and the suture were stretched, while the plate
was compressed against the bone. Initially, with an increase in the applied displacement, the load increase was almost linear. The slope of the linear portion of the elongation–force curve was assumed to represent the compliance of the tendon–suture–bone system (Table 1). The statistical analysis showed no difference between the mean compliance values of the four groups (A N O V A , P > 0:05). A further increase in the distracting displacement produced a gap larger than 10 mm between the tendon and the bone, which was observed before reaching the maximum tensile strength in all investigated repairs. Failure of the repair always occurred at the stitch–tendon interface. Two failure modes were observed: rupture of the suture or pull out of the suture from the tendon. Of the 14 repairs performed using the Modified S.C.O.I. grasping technique, 6 failed by failure of the thread at the stitch level and the other 8 failed by suture pull out of the tendon. All the 14 repairs performed using the Modified S.C.O.I. grasping technique failed by suture pull out of the tendon. Fig. 7D shows the tensile strength measured for the four investigated configurations. The plate used for bone augmentation had no statistical effect (A N O V A , P > 0:05) on the tensile strength of the repair. In addition, no statistical difference (A N O V A , P > 0:05) was found between the mean values of the tensile strength calculated for the two tendon grasping techniques, in spite of the different modes of failure observed.
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
727
Table 1 Compliance of the entire repair construct calculated for the linear part of the load–displacement curve Configuration Bone quality Simulated Simulated Simulated Simulated
osteoporotic osteoporotic osteoporotic osteoporotic
bone bone bone bone
Grasping technique
Button plate type
Two Two Two Two
BP7H BP7H PS2H PS2H
Mod. Mod. Mod. Mod.
Mason–Allen S.C.O.I. Mason–Allen S.C.O.I.
4. Discussion The present study investigates the mechanical properties of different techniques for rotator cuff repair. The proposed experimental method used an animal model, testing the infraspinatus tendon and humerus of sheep. The sheep model has already been used by others to investigate the mechanical strength of different rotator cuff repair techniques. In fact, the histological and mechanical characteristics of sheep infraspinatus tendon are similar to those of human supraspinatus tendons (Gerber et al., 1994; Lewis et al., 2001). It is recognised that there are a number of limitations in this study. The experimental method cannot simulate degeneration of the tendon that may be observed in delayed repair in human. All the samples (tendon-humerus) were collected intact and, at the time of testing, the tendon was transected to represent an acute tear. While this does not represent an actual tear, the method of transection was preferred, being repeatable for every specimen. In addition, the sheep humera were not osteoporotic. The cortical bone of the greater tuberosity was removed to simulate the effect of osteoporosis. Again, this procedure was used to ensure repeatability. Finally, the experimental setup did not simulate the physiological activities that can stress the repair in vivo. Therefore, the absolute values of the repair tensile strength found in this study cannot be directly transferred to clinically relevant numbers. Nevertheless, the experimental data collected can be useful to highlight the difference between various rotator cuff repair techniques. One purpose of the present study was to compare the holding power of three grasping techniques: the Modified S.C.O.I., the Modified Mason–Allen, and the Mattress. The failure mode of the tendon–suture system was identified as suture failure or tendon cut through. The experimental data showed that the Mattress group had a lower load to failure than the remaining two groups. Out of the 10 tendon–suture constructs of the Mattress group, the suture cut through the tendon in 8. Conversely, the Modified S.C.O.I. and the Modified Mason–Allen did not slip within the tendon. In all cases, failure occurred for suture breakage at the thread knot. Hence, the holding power of the Modified Mason–Allen stitch and of the Modified S.C.O.I. stitch is larger than that measured for the Mattress stitch.
Entire repair compliance (mm/N) Mean (S.D.) 0.046 0.048 0.045 0.047
(0.013) (0.009) (0.014) (0.011)
The recurrence of tendon–suture construct failure at the thread knot level suggests that the knot is a weak link of the suture loop. Gerber et al. (1994) reported an ultimate tensile strength of 82 ± 3 N for a single Ethibond no. 2 thread. In the present study the suture loop tied together by a single knot failed at about 120 N. Since the load was equally distributed on the two threads of the suture loop (the load was equally distributed over the pulley), it can be calculated that breakage of the suture occurred at about 60 N. This reduced value of the tensile strength confirms that the knot weakened the thread, as already suggested by Burkhart et al. (1996). Two different reasons explain this observation: the knot caused a stress concentration in the thread, the thread was damaged at the knot level during tying. To improve this weak link of the suture loop, the Ôdouble knotÕ was introduced. This solution showed a tensile strength over twice that of the Ôsingle knotÕ (252 and 119 N, respectively) but the Ôdouble knotÕ involve fixing two sutures. No significant difference was found in the tensile strength values of the Ôdouble knotÕ and of two Ôsingle knotsÕ working in parallel. Hence, the Ôdouble knotÕ does not improve the tensile strength of the suture locking. Additionally, the Ôdouble knotÕ may require a greater effort by the surgeon to reduce the gap between the tendon and the bone surface. In fact, knotting two stitches simultaneously implies that a wider portion of the contracted tendon needs to be pulled to bone than that in case of a Ôsingle knotÕ. Nevertheless, the double knot seems preferable because this solution would reduce differences in the load stressing the two suture loops tied together and finally should distribute the load on the tendon more homogeneously. The results of the test series on suture-to-bone fixation showed that fixation over augmented bone was superior than that over unaugmented bone when the cortical bone has poor mechanical characteristics (osteoporotic bone). These findings are in agreement with previous reports (Sward et al., 1992; Caldwell et al., 1997). The metal plate increases the contact area between the repair and the bone reducing the compressive pressure on the bone and finally the risk of bone failure. Conversely, the bone augmentation seems mechanically useless when fixation is performed on good quality cortical bone. In fact, in the present investigation good quality cortical bone withstood compressive pressure
728
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729
until suture failure occurred, with negligible tissue damage at the contact area. It is surprising that in this test series the Ôdouble knotÕ failed at a higher load than that measured in testing the knot suture alone. Theoretically, the strength of the repair without tendon should not be higher than that of the Ôdouble knotÕ. This observation suggests the hypothesis that the load is partially transmitted from the suture to the bone within the tunnels by friction. Therefore, the knot (the weak point of the suture loop), being located on the far side of the tunnels, is only stressed by a fraction of the load stressing the threads. Thus, higher loads can be applied to the suture before reaching the ultimate tensile strength of the knot. For all of the investigated entire repair constructs, the bone–suture fixation involved a Ôdouble knotÕ tied over a plate augmentation. The entire repairs never failed over the metallic plate or at the suture knot level. Hence, the bone augmented fixation technique transferred the weakest point of the repair to the suture–tendon interface or at the suture material. These experimental results are in agreement with previous findings (Gerber et al., 1994; Sward et al., 1992). The elongation under tensile load was not different for the investigated repair techniques. The large compliance of the four entire repairs was unexpected. A large gap between the tendon end and the bone surface of up to 10 mm was observed prior to failure of the repair. Hence, physiological loads may generate a gap at the tendon–bone interface, where values of up to 300 N have been reported in the supraspinatus tendon for abduction of the arm (Wallace, 1984). The findings of Gerber et al. (1999) and Burkhart et al. (1997a,b) support this statement. In testing the entire repair, the two investigated grasping techniques, the Modified Mason–Allen stitch and the Modified S.C.O.I., showed a comparable tensile strength. However, two failure modes of the tendon– suture system were identified: rupture of the suture at the stitch level or pull out of the suture from the tendon. In all cases, the Modified S.C.O.I. repair failed by cutting through the tendon. Of the 14 Modified Mason– Allen repairs tested, 8 failed by cutting through the tendon, and the other 6 failed by suture breakage at the stitch level. These findings would suggest that the Modified Mason–Allen stitch has a higher tendon holding power than the Modified S.C.O.I. stitch, but at the same time this results in a higher stress level in the suture material for the Modified Mason–Allen stitch in comparison with stress level for the Modified S.C.O.I. stitch. Both the holding power and the stress level in the suture material are an intrinsic characteristic of the stitch affected by the tendon tissue quality and by thread size. The surgeon cannot change the tissue quality of the patient tendon. Conversely, the suture size can be increased with no need for larger transosseous tunnels. Considering the Modified Mason–Allen stitch strength
from this study and the results of Gerber et al. (1994), it can be seen that an increase in one size of the suture can increase the tensile strength by 50%. The compliance of the repair would also decrease. Hence, it would be recommended to use a suture size larger than no. 2. However, a limit exists in increasing the size of thread: too larger a suture may promote local irritation due to possible contact with the acromion. In addition, a tendon–suture augmentation would improve the tensile strength of the repair (France et al., 1989). Nevertheless, adding an augmentation patch at the tendon–thread interface may cause impingement (Sward et al., 1992).
5. Conclusions The use of ÔreinforcedÕ stitches (Modified Mason– Allen stitch or Modified S.C.O.I. stitch) for tendon grasping seems recommendable due to the higher tensile strength of these stitches in comparison with the Mattress stitch. The Ôdouble knotÕ locking, reducing a possible unequal load distribution between two suture loops, seems preferable to two Ôsingle knotsÕ working in parallel. The suture-to-bone fixation over augmented cortical bone had a higher mechanical strength than fixation over osteoporotic bone. Conversely, augmentation over good quality cortical bone seemed to produce a negligible improvement of the suture-to-bone fixation strength. Nevertheless of the efforts to improve the surgical technique and finally the strength of the repair, the compliance of the repair remain a problem. Post-operative protection of the repaired rotator cuff from tension load is necessary.
Acknowledgements The authors wish to thank Luigi Lena for the illustrations, Roberta Fognani for technical support during the experiments, Cristina Ancarani for the statistical analysis and Victor Waide for help with the English revision.
References Adamson, G.J., Tibone, J.E., 1993. Ten-year assessment of primary rotator cuff repairs. J. Shoulder Elbow Surg. 2, 57–63. Armitage, P., Berry, G., 1994. Statistical Methods in Medical Research, third ed. Blackwell Scientific Publications, Oxford. Becker, B.J., 1988. Synthesizing standardized mean-change measures. Brit. J. Math. Stat. Psychol. 41, 257–278. Bigliani, L.U., Kimmel, J., McCann, P.D., Wolfe, I., 1992. Repair of rotator cuff tears in tennis players. Am. J. Sports Med. 20, 112–117. Burkhart, S.S., Fischer, S.P., Nottage, W.M., Esch, J.C., Barber, F.A., Doctor, D., et al., 1996. Tissue fixation security in transosseous
M. Baleani et al. / Clinical Biomechanics 18 (2003) 721–729 rotator cuff repairs: a mechanical comparison of simple versus mattress sutures. Arthroscopy 12, 704–708. Burkhart, S.S., Diaz Pagan, J.L., Wirth, M.A., Athanasiou, K.A., 1997a. Cyclic loading of anchor-based rotator cuff repairs: confirmation of the tension overload phenomenon and comparison of suture anchor fixation with transosseous fixation. Arthroscopy 13, 720–724. Burkhart, S.S., Johnson, T.C., Wirth, M.A., Athanasiou, K.A., 1997b. Cyclic loading of transosseous rotator cuff repairs: tension overload as a possible cause of failure. Arthroscopy 13, 172–176. Caldwell, G.L., Warner, J.P., Miller, M.D., Boardman, D., Towers, J., Debski, R., 1997. Strength of fixation with transosseous sutures in rotator cuff repair. J. Bone Joint Surg Am. 79, 1064–1068. Cofield, R.H., Parvizi, J., Hoffmeyer, P.J., Lanzer, W.L., Ilstrup, D.M., Rowland, C.M., 2001. Surgical repair of chronic rotator cuff tears. A prospective long-term study. J. Bone Joint Surg Am. 83-A, 71–77. Cordasco, F.A., Bigliani, L.U., 1997. The rotator cuff. Large and massive tears. Technique of open repair. Orthop. Clin. North Am. 28, 179–193. Craft, D.V., Moseley, J.B., Cawley, P.W., Noble, P.C., 1996. Fixation strength of rotator cuff repairs with suture anchors and the transosseous suture technique. J. Shoulder Elbow Surg. 5, 32–40. France, E.P., Paulos, L.E., Harner, C.D., Straight, C.B., 1989. Biomechanical evaluation of rotator cuff fixation methods. Am. J. Sports Med. 17, 176–181. Gazielly, D.F., Gleyze, P., Montagnon, C., 1994. Functional and anatomical results after rotator cuff repair. Clin. Orthop. 304, 43– 53. Gerber, C., Schneeberger, A.G., Beck, M., Schlegel, U., 1994. Mechanical strength of repairs of the rotator cuff. J. Bone Joint Surg. Br. 76, 371–380. Gerber, C., Schneeberger, A.G., Perren, S.M., Nyffeler, R.W., 1999. Experimental rotator cuff repair. A preliminary study. J. Bone Joint Surg Am. 81, 1281–1290. Gerber, C., Fuchs, B., Hodler, J., 2000. The results of repair of massive tears of the rotator cuff. J. Bone Joint Surg Am. 82, 505–515. Goradia, V.K., Mullen, D.J., Boucher, H.R., Parks, B.G., OÕDonnell, J.B., 2001. Cyclic loading of rotator cuff repairs: a comparison of bioabsorbable tacks with metal suture anchors and transosseous sutures. Arthroscopy 17, 360–364. Harryman, D.T., Mack, L.A., Wang, K.Y., Jackins, S.E., Richardson, M.L., Matsen, F.A., 1991. Repairs of the rotator cuff. Correlation of functional results with integrity of the cuff. J. Bone Joint Surg Am. 73, 982–989. Hata, Y., Saitoh, S., Murakami, N., Seki, H., Nakatsuchi, Y., Takaoka, K., 2001. A less invasive surgery for rotator cuff tear: mini-open repair. J. Shoulder Elbow Surg. 10, 11–16. Hattrup, S.J., 1995. Rotator cuff repair: relevance of patient age. J. Shoulder Elbow Surg. 4, 95–100.
729
Hawkins, R.J., Misamore, G.W., Hobeika, P.E., 1985. Surgery for full-thickness rotator-cuff tears. J. Bone Joint Surg. Am. 67, 1349– 1355. Hecker, A.T., Shea, M., Hayhurst, J.O., Myers, E.R., Meeks, L.W., Hayes, W.C., 1993. Pull-out strength of suture anchors for rotator cuff and Bankart lesion repairs. Am. J. Sports Med. 21, 874–879. Iannotti, J.P., Bigliani, L.U., 1998. The Rotator Cuff: Current Concepts and Complex Problems. American Academy of Orthopaedic Surgeon, Rosemont. Jost, B., Pfirrmann, C.W., Gerber, C., Switzerland, Z., 2000. Clinical outcome after structural failure of rotator cuff repairs. J. Bone Joint Surg Am. 82, 304–314. Knudsen, H.B., Gelineck, J., Sojbjerg, J.O., Olsen, B.S., Johannsen, H.V., Sneppen, O., 1999. Functional and magnetic resonance imaging evaluation after single-tendon rotator cuff reconstruction. J. Shoulder Elbow Surg. 8, 242–246. Lewis, C.W., Schlegel, T.F., Hawkins, R.J., James, S.P., Turner, A.S., 2001. The effect of immobilization on rotator cuff healing using modified Mason–Allen stitches: a biomechanical study in sheep. Biomed. Sci. Instrum. 37, 263–268. McKee, M.D., Yoo, D.J., 2000. The effect of surgery for rotator cuff disease on general health status. Results of a prospective trial. J. Bone Joint Surg. Am. 82-A, 970–979. McLaughlin, H.L., 1944. Lesion of the musculotendinous cuff of the shoulder. The exposure and treatment of tears with retraction. J. Bone Joint Surg. Am. 26, 31–51. Reed, S.C., Glossop, N., Ogilvie-Harris, D.J., 1996. Full-thickness rotator cuff tears. A biomechanical comparison of suture versus bone anchor techniques. Am. J. Sports Med. 24, 46–48. Romeo, A.A., Hang, D.W., Bach, B.R., Shott, S., 1999. Repair of full thickness rotator cuff tears. Gender, age, and other factors affecting outcome. Clin. Orthop. 367, 243–255. Rossouw, D.J., McElroy, B.J., Amis, A.A., Emery, R.J., 1997. A biomechanical evaluation of suture anchors in repair of the rotator cuff. J. Bone Joint Surg. Br. 79, 458–461. Sward, L., Hughes, J.S., Amis, A., Wallace, W.A., 1992. The strength of surgical repairs of the rotator cuff. A biomechanical study on cadavers. J. Bone Joint Surg. Br. 74, 585–588. Thomazeau, H., Boukobza, E., Morcet, N., Chaperon, J., Langlais, F., 1997. Prediction of rotator cuff repair results by magnetic resonance imaging. Clin. Orthop. 344, 275–283. Wallace, W.A., 1984. Evaluation of the force, the I.C.R. and the neutral point during abduction of the shoulder. In: 30th Annual Meeting of the Orthopaedic Research Society, Atlanta, Georgia. Worland, R.L., Arredondo, J., Angles, F., Lopez_Jimenez, F., 1999. Repair of massive rotator cuff tears in patients older than 70 years. J. Shoulder Elbow Surg. 8, 26–30. Wulker, N., Melzer, C., Wirth, C.J., 1991. Shoulder surgery for rotator cuff tears. Ultrasonographic 3-year follow-up of 97 cases. Acta Orthop. Scand. 62, 142–147.