In-vitro comparison of 3 knotting techniques for lateral fabellotibial suture stabilization

Abstract

This study evaluated the biomechanical characteristics of a single self-locking knot (sSLK) and a double self-locking knot (dSLK) compared with the square knot (SQ) for stabilization of cranial cruciate ligament rupture. Each knot underwent monotonic tensile and cyclical loading. Starting tension, elongation, stiffness, and load to failure were all evaluated. A value of P < 0.05 was considered significant. Starting tension, overall stiffness, and load to failure were all significantly greater in both the sSLK and dSLK compared with the SQ. There was no difference in elongation among the knots. There were no significant differences in starting tension, elongation, stiffness, and load to failure between the sSLK and the dSLK. The self-locking knots were stronger and stiffer than the SQ; there is no biomechanical advantage in using the dSLK compared with the sSLK.

Introduction

Cranial cruciate ligament (CCL) rupture is one of the most common orthopedic conditions encountered in dogs (1–4). In 2003, it was estimated that over 1 billion US dollars were spent for the treatment of cranial cruciate ligament rupture in the United States (5). The CCL functions to prevent hyperextension of the stifle, internal rotation, and cranial translation of the tibia. Surgical correction is recommended in dogs > 15 kg (6) or those with suspected meniscal tears. Vasseur (7) found that 75% of dogs < 15 kg had apparent resolution of lameness with recovery taking an average of 4 mo, but that 25% of medically treated dogs still required some form of stabilization. This is longer than what is expected with surgical intervention (1,2,8). Resolution of clinical signs can be achieved with medical therapy, but continued instability may lead to rapid progression of osteoarthritis, meniscal tears, and subclinical pain.

Techniques designed to stabilize the CCL deficient stifle can be categorized as intracapsular, extracaspsular, and osteotomies [i.e., tibial plateau leveling osteotomy (TPLO), tibial tuberosity advancement (TTA), and cranial tibial wedge osteotomy (CTWO)] with the overall goal being to restore normal stifle biomechanics and stop the progression of osteoarthritis (OA). To date no surgical technique has been described that accomplishes these goals completely (7,9–12).

Extracapsular suture stabilization was described by DeAngelis and Lau in 1970 (13). Since the original description others have tried to find the optimal material and fixation method (e.g., suture material, knotting versus crimping, and using bone anchors) (14,15) to restore normal stifle biomechanics. The cranial tibial shearing force should be eliminated, as this would allow postoperative return of limb function as well as possibly stop and/or eliminate the progression of OA. All materials used to stabilize the stifle via the extracapsular method have the possibility of failing, which would leave stability to be maintained by mechanisms such as periarticular fibrosis and muscular strength. Causes of failure include premature failure of the suture, most commonly at the knot (16), and alteration in the attachment points (17–19). The ideal material should have no creep, be biologically inert, aseptic, easily handled, inexpensive, and provide excellent knot security, knot compactness, and the ability to withstand cyclical and tensile loading (20).

Knotting is a point of stress, which may adversely affect the biomechanical properties during times of high tension on the suture loop, leading to failure of the knot (21). Furthermore, large diameter monofilament nylon leader line (NLL) can leave a bulky knot that is hard to tie securely, which may increase patient morbidity and the incidence of knot slippage. Also, NLL may undergo significant elongation (22), which can lead to cranio-caudal stifle translation and is detrimental to long-term stability of the stifle (23). However, NLL has several advantages, including availability in several sizes, and being inexpensive, strong, biologically inert, and easy to handle.

In 1999 McKee and Miller described the use of a self-locking knot (SLK) (24). The purpose was to allow the unassisted surgeon to create a secure knot without the loss of loop tension. The SLK works because the first throw of the knot creates a locking-loop effect. The procedure was described with either a single loop (sSLK) or a double loop (dSLK). It was speculated that in dogs > 35 kg the single loop could break; therefore, a double loop was advisable. To our knowledge no in vitro biomechanical studies or in vivo clinical studies have been performed to compare single versus double loops. Previous studies have evaluated monotonic loading comparing the SLK with other knots (16,25); however, cyclic loading is considered to be more clinically applicable. The authors are aware of only 1 study in which cyclic loading was evaluated using the sSLK; however, no data exist regarding cyclic loading of the dSLK (14). According to McKee and Miller (24) the dSLK should have superior biomechanical characteristics (overall strength, stiffness, and elongation) compared with both the sSLK and the square knot (SQ). This could be explained because the dSLK may provide increased friction. However, with the added suture material there is the possibility of greater knot bulk leading to greater patient morbidity with the dSLK.

Previous studies (14,16,26) have shown the SLK can easily withstand the tension needed to provide adequate stabilization in vitro; however, there is still the question as to whether the dSLK is stronger than the sSLK, and whether elongation or stiffness is different between the two. Our hypothesis is that there will be no biomechanical advantage using the dSLK compared with the sSLK. Both the dSLK and sSLK will exhibit greater load to failure, greater stiffness, and less elongation than the SQ.

Materials and methods

Twenty-seven kilogram monofilament NLL (Securos company, Fiskdale, Massachusetts, USA) was used during testing of the sSLK, dSLK, and SQ techniques. Loops were fixed by the same investigator (D.D.) using gloved hands, around smooth metal hooks 30 mm apart (Figure 1) to ensure a consistent size of the loop. Loop diameter was chosen based on clinical relevance. The square knot had tension maintained at 100 N by the same assistant (S.E.) distracting each end of the suture. Square knots were formed by clamping the first throw with Mayo-Hagar needle holding forceps, while a second throw was placed; 3 additional throws were then applied. The SLK was formed as described by McKee and Miller (24) utilizing either a single or double loop unassisted (Figures 2 and 3). The free ends of the NLL were passed through the loop (sSLK) from lateral to medial (Figure 2). For the dSLK the loop was folded over itself to create 2 loops (Figure 3). Proximocaudal traction on the 2 ends reduces the size of the loop(s). The direction of the traction is then reversed (distocranially). The proximocaudal and distocranial traction were repeated to achieve a tension of 100 N. Four additional throws were then applied after the creation of the SLK (24). The ends of all loops were cut approximately 3 mm from the knot.

Figure 1.

Machine testing set-up, illustrating smooth metal hooks 30 mm apart.

Figure 2.

Placement of the sSLK as described by McKee and Miller (24).

Figure 3.

Placement of the dSLK as described by McKee and Miller (24). Note the double loop where the loop has been folded on itself.

Monotonic loading

Fifteen knots (n = 15 per group) were tied for each of the 3 knots and were subjected to monotonic tensile loading as detailed in a previous report (20). Each knot was subjected to loading at 300 mm/min until failure (Instron Universal Testing Instrument Model 1011, Instron, Norwood, Massachusetts, USA).

Cyclic loading

Eight knots (n = 8 per group) from each of the 3 knots underwent cyclical loading (Instron Universal Testing Instrument Model 1011). Loops were pretensioned to 10 N and then subjected to ramp loading and unloading at 200 N/min. Every 5 cycles the maximal force was increased by 100 N, starting at 100 N. The minimum (valley) force to which the loop was returned at the end of the cycle was 10 N.

Data analysis

In monotonic testing the starting tension at completion of tying (N), peak load to failure (N), peak elongation (mm) and stiffness (N/mm) were recorded. Stiffness was calculated as the slope of the best-fit line through the linear portion of the loading curve (force versus displacement). For cyclic loading peak load to failure (N), peak elongation (mm), number of cycles to failure, and stiffness (N/mm) were evaluated. During cyclic loading stiffness was calculated as the peak-to-peak force divided by the peak-to-peak displacement. Method of failure (slippage or breakage) was recorded for both monotonic and cyclic loading. Maximal elongation was determined from the end of the linear region of the force versus displacement curve. In this study elongation will refer to the area of elastic deformation with maximal elongation occurring at the yield point. In vivo, catastrophic failure will result after the yield point prior to the plastic phase of deformation.

Statistical analysis

For a P-value of 0.05 to indicate significance and a power of 0.95, 18 loops (n = 6 per group) would need to be tested in both monotonic and cyclic loading. Analysis of variance was performed using the procedures of statistical analysis system (SAS 9.2, SAS Institute, Cary, North Carolina, USA). The significance of treatment effect (SQ, sSLK, dSLK) and differences of treatment means were determined using PROC GLM procedure which uses the method of least squares to fit general linear models. Correlations were determined using PROC CORR which computes Pearson correlation coefficients, 3 nonparametric measures of association, and the probabilities associated with each statistic. For all statistical analyses, values of P < 0.05 were considered significant.

Results

All the loops in both the monotonic and cyclic tests failed by breakage at the knot or within 2 to 4 mm of the final location of the knot for all 3 knots. All values, reported as the mean ± standard deviation (SD) are listed in Tables 1 and 2. Starting tension with the SQ was significantly decreased compared with both the sSLK and the dSLK where no difference was detected. The tension noted just after the knot was tied was approximately 35 to 40 N with the sSLK and dSLK and < 20 N with the SQ (P < 0.0001). In monotonic tensile loading the sSLK and dSLK were significantly stiffer than the SQ (P < 0.0001). The sSLK achieved the highest stiffness of 132 N/mm; however, this was not significantly different from the dSLK. Both the sSLK and dSLK achieved a significantly greater load to failure than the SQ (P < 0.0001), with sSLK achieving the greatest load to failure (801 N). Elongation was not significantly different among the knots, with an average elongation of 10.1 ± 0.26 mm.

Table 1.

Biomechanical data reported as the mean ± SD for all 3 knots for monotonic data

Knot	Starting tension (N)	Elongation (mm)	Stiffness (N/mm)	Load to failure (N)
SQ	18.7 ± 7.0	10.2 ± 3.6	64.3 ± 12.5	333.9 ± 86.0
sSLK	37.7 ± 7.2	9.8 ± 0.9	132.7 ± 9.6	801.8 ± 34.8
dSLK	39.9 ± 16.6	10.3 ± 1.8	125.5 ± 27.9	743.9 ± 80.5

N — Newton, SQ — square knot, sSLK — single self-locking knot, dSLK — double self-locking knot.

Table 2.

Biomechanical data reported as the mean ± SD for the 3 knots for cyclic data

Knot	Elongation (mm)	Stiffness (N/mm)	Load to failure (N)	Number of cycles until failure
SQ	5.99 ± 2.5	98.1 ± 14.3	443.8 ± 74.8	20.6 ± 4.1
sSLK	10.25 ± 1.1	155.3 ± 8.7	951.8 ± 95.0	44.3 ± 4.5
dSLK	8.19 ± 1.8	145.5 ± 15.0	941.9 ± 132.2	40.4 ± 4.5

N — Newton, SQ — square knot, sSLK — single self-locking knot, dSLK — double self-locking knot.

For cyclic tensile loading the overall stiffness was significantly greater for both the sSLK and dSLK compared with the SQ (P < 0.0001); sSLK achieved the greatest stiffness of 155 N/mm. Load to failure with cyclic loading was significantly greater with both the sSLK and dSLK compared with the SQ (P < 0.0001). Elongation was the same among the knots, with an average of 8.0 ± 2.1 mm. The number of cycles that resulted in failure was significantly greater with both the sSLK and dSLK compared with the SQ (P < 0.0001). For both monotonic and cyclic tensile loading there were no significant differences between the sSLK and dSLK.

Discussion

In this in vitro study we demonstrated that there is no added benefit in using the dSLK over the sSLK. As expected, the sSLK and dSLK were biomechanically stronger than the SQ because the SLK is composed of 2 strands while the SQ has only 1. Our results are in close agreement with others in the starting tension achieved (14,27,28). We found that both the sSLK and the dSLK have double the amount of tension applied to the knot unassisted compared with the assisted SQ. Also, the tension achieved is comparable to that with the single strand crimping system (14).

Tonks et al (19) found that lateral compartment contact pressures could be altered when over-tightening the extra capsular suture; however, optimal tension could not be determined. They found that tension in the unloaded stifle greater than 40 N may change the lateral contact pressures. Our study found tensions of 37 N and 39 N were reached with the sSLK and dSLK, respectively. This is under the 40 N found in the previous study but close enough that caution should be used when tying the self-locking knot as to not over-tighten the suture. Having suture under too much tension could possibly over-constrain the joint; a loss of tension would result in joint instability. Furthermore, tightening the suture at different joint angles or not using isometric points will have different effects on the starting tension.

Elongation as reported by others (14,21) should be a primary consideration when choosing not only suture material but also the type of knot when performing an extracapsular repair. The reasoning behind this is that many surgeons accept 2 mm of cranial drawer post-operatively (20). Elongation contributes to cranio-caudal stifle translation and can be detrimental to long-term stifle stability (23). Our results indicated no difference in elongation between any of the knots. For both monotonic and cyclic testing, elongation was between 8.0 and 10.0 mm. Another study that evaluated the sSLK using 60 lb nylon leader line reported an elongation of 14.8 mm with the sSLK and 10.4 mm with the square knot (16). These data should be interpreted with caution as it is unlikely that a canine tibia would be displaced 10 or more mm relative to the femur. Caporn and Roe (27) found that as nylon suture increases in size it becomes more difficult to tighten each throw of the knot. With progressive loading, the small gaps between each throw may tighten which would lead to loop elongation (28,29). Elongation can result from not only tightening the knot but also elongation of the material itself while undergoing testing. Mulon et al (26) found the elongation of USP2 Polydioxanone used with the sSLK to be 9.37 mm at failure.

Stiffness when calculated by dividing force by elongation helps to express the relationship of strength to elongation and can be used to make comparisons among materials and knotting methods. For both monotonic and cyclic testing the stiffness was significantly greater for the sSLK and dSLK than for the SQ by roughly 2-fold, which is expected given that the SLK utilizes 2 strands of suture while the SQ only has 1. Stiffness is defined as a material’s ability to resist deformation under load. Wingfield (30) reported that normal stifle stiffness in the canine cranial cruciate ligament ranges from 148 to 224 N/mm. Our results with monotonic loading indicated stiffness of 132 N/mm and 125 N/mm for sSLK and dSLK, respectively, while the SQ achieved an overall stiffness of 64 N/mm, which is similar to the results of Vianna and Roe (14).

During cyclic testing the sSLK had a stiffness of approximately 155 N/mm, the dSLK had an overall stiffness of 145 N/mm, and the SQ had a stiffness of 98 N/mm. There are no studies evaluating cyclic biomechanical testing between the sSLK and dSLK. In 2006 Vianna and Roe (14) evaluated the sSLK with cyclic loading, reporting only the starting tension and loop elongation, not stiffness. Cabano (25) tested an interlocking loop construct with a crimp and recorded loop tension and elongation with cyclic loading. Burgess et al (20) found the mean stiffness in cyclic loading using nylon leader line with a square knot and a crimp system to be approximately 100 N/mm and 150 N/mm, respectively, which is very similar to our results.

In recommending material and type of knot to repair a ruptured cruciate ligament the material should achieve 10% to 25% of the strength of the ligament at failure (31). The normal canine cruciate ligament can withstand forces of up to 2130 N but, under normal physiologic loads, it is subjected to 400 to 600 N (27,30). Therefore a repair technique should be able to withstand approximately 160 to 400 N. In monotonic loading the load to failure was 800 N for sSLK and 740 N for dSLK; the SQ had a load to failure of 333 N. These correspond to the report by Peycke et al (16) who found a load to failure of 760 N for sSLK and a load to failure of 331 N with the square knot and 27-kg nylon leader line. The cyclic data showed a load to failure of around 940 to 950 N for both sSLK and dSLK while the SQ had a load to failure of about half that. Vianna and Roe (14) reported that each of their SLK constructs was still intact at 850 N, but did not report ultimate failure loads. The cyclic data using the self-locking knot is comparable to both the crimping system and fiber tape at 767 N and 861 N, respectively (14,20).

Since points of stress such as sharp bends could create a weak point, several studies have evaluated different types of knots and crimping systems. Burgess et al (20) found that crimped loops were weaker than knotted loops except for fiber wire. These findings were similar to other studies with nylon leader (16,22,32). The self-locking knot may therefore be advantageous compared with the crimp system. In contrast, Vianna and Roe (14) suggest that a crimped system is mechanically superior to knotted loops because it resists loading more effectively before becoming permanently elongated. Failure of the knot is a common reason why persistent stifle instability may exist after an extra capsular repair. Friction, internal interference, and the presence of slack have all been identified as contributing factors in knot security (20). It has been recommended that 5 throws be used to secure the knot and eliminate slippage when using nylon leader line (27); however, a greater number of knots may cause tissue irritation. In our study, all the self-locking knots broke at or very near the knot with no slippage noted. This further demonstrates the ability of the self-locking knot to lock after the first throw thus possibly preventing or eliminating slippage.

Limitations to our study include the in vitro nature of the work; caution should be used when extrapolating our results to in vivo performance. Our testing protocol does not take into account all the possible forces acting on the stifle joint, isometry of attachment points, or additional stabilization contributed by the surrounding soft tissues. Monotonic testing provides baseline information for biomechanical comparisons but has little clinical value. Cyclic testing tries to simulate in vivo conditions by placing the suture through various cycles until failure. This provides better clinical representation than monotonic testing; however, the number of cycles used in our testing is far less than what is expected for suture in vivo. The results in terms of load to failure and elongation reached in vitro far exceed what are expected to be achieved in vivo. Another limitation to our study is that the data reported for the SQ should not be directly compared to that for the SLK as this was not the aim. The aim was to compare the biomechanical characteristics of the sSLK versus the dSLK while the SQ was tested to validate our testing method. Lastly, the hooks were 30 mm apart leaving only 15 mm between the hook and knot, suggesting that some of the sutures may have actually failed at the hooks.

In conclusion, we demonstrated that there is no advantage in using the dSLK compared with the sSLK. Therefore, we do not recommend using the dSLK as it has no biomechanical advantage in vitro and it creates added bulk, which may lead to further tissue irritation. As previously shown, the self-locking knot is superior to the square knot in monotonic testing (16,26), and in cyclic testing (14); this is expected given that the SQ has only 1 strand of suture while the SLK has 2. Superior initial tension can be achieved with the SLK, without assistance, similar to what is accomplished with a crimping system (14). CVJ