INTRASYNOVIAL FLEXOR TENDON REPAIR: A BIOMECHANICAL STUDY OF VARIATIONS IN SUTURE APPLICATION IN HUMAN CADAVERA

Abstract

To improve the functional outcomes of intrasynovial tendon suture, prior experiments evaluated individual technical modifications used in the repair process. Few studies, however, have assessed the combinatorial effects of those suture modifications in an integrated biomechanical manner, including a sample size sufficient to make definitive observations on repair technique. 256 flexor tendon repairs were performed in cadavera, and biomechanical properties were determined. The effects of five factors for flexor tendon repair were tested: core suture caliber (4-0 or 3-0), number of sutures crossing the repair site (4- or 8-strand), core suture purchase (0.75 cm or 1.2 cm), peripheral suture caliber (6-0 or 5-0), and peripheral suture purchase (superficial or 2 mm). Significant factors affecting the properties of the repair were the number of core suture strands and the peripheral suture purchase. The least significant factors were core suture purchase and peripheral suture caliber. The choice of core suture caliber affected the properties of repair marginally. Based on these results, we recommend that surgeons continue to focus on multi-strand repair methods, as the properties of 8-strand repairs were far better than those of 4-strand repairs. To resist gap formation and enhance repair strength, a peripheral suture with 2mm purchase is also recommended. Finally, since core suture caliber affected some biomechanical properties, including the failure mode, a 3-0 suture could be considered, provided that future in vivo studies can confirm that gliding properties are not adversely influenced.

INTRODUCTION

A substantial number of upper extremity injuries include lacerations to the intrasynovial flexor tendons of the hand.¹ Although significant advances have been made in the treatment of tendon transections, the formation of repair-site adhesions and the development of repair site gap formation leading to tendon rupture have led to highly variable clinical outcomes.^{2, 3} Surgeons are confronted with many choices when repairing intrasynovial flexor tendon lacerations, among them the caliber and quantity of suture used as a core construct, the depth to which the peripheral epitendinous suture is passed, the distance from the cut tendon end that the epitendinous tendon is introduced, and the caliber of the epitendinous suture itself. These variables have been evaluated independently in in vivo and ex vivo models so that appropriate recommendations could be made regarding their clinical effect.^4-11 Despite these studies, it remains unclear which combination of operative technique modifications is most helpful in resisting gap formation and providing sufficient strength to allow early rehabilitation. For example, if the caliber of the core suture were 3-0, and an 8-strand repair were utilized, is the caliber of the epitendinous suture or the distance/depth to the repair site inconsequential? Similarly, if the tendon is of a size or thickness at the laceration that makes placement of a higher caliber core suture impossible, can utilization of a higher caliber epitendinous suture obviate the negative effects of the smaller core suture? While the surgeon may approach the repair with the intuitive concept that more suture is better, the application of this principle may be difficult due to anatomical considerations.

Most of our knowledge regarding surgical factors in flexor tendon repair was established in studies comparing one or two isolated variables. From these studies, several important factors were identified, including the number of core sutures⁴, the caliber of the core suture^{5, 12}, the purchase of the core suture^{7, 13}, and the purchase of the epitenon (i.e., “peripheral”) suture^{6, 14}. Increases in each of these factors individually produced increases in strength, stiffness, and resistance to gap formation. However, a comparison of their combined effects on mechanical properties of time-zero repairs is lacking. In addition, since much of our knowledge is based on investigating a single variable, interactions among variables are unknown. In the current study, we performed a systematic biomechanical examination of 5 of the most commonly studied factors: core suture caliber, core suture strands, core suture purchase, peripheral suture caliber, and peripheral suture purchase. To determine the impact of each variable on the biomechanical characteristics of midsubstance tendon suture, the factors were varied in cadaveric flexor tendon repairs. We hypothesized that the number of core suture strands and the caliber of the core suture would have the greatest effects. Although many variables influence clinical success (e.g., strength, gliding, repair-site bulk), we focused on initial strength, stiffness, and resistance to gapping, as these properties dictate the risk for rupture and the ability to perform rehabilitation.

MATERIALS AND METHODS

Research design

256 intrasynovial flexor tendon lacerations followed by surgical repair were performed in human cadavers. The repairs were tested in tension to determine biomechanical properties. The aggregate effects of 5 independent factors and their combinatory effects were evaluated: core suture caliber (4-0 or 3-0), number of core suture strands crossing the repair site (4- or 8-strand), core suture purchase (0.75 or 1.2 cm), peripheral suture caliber (6-0 or 5-0), and peripheral suture purchase (superficial or 2 mm). Demographic information was unavailable for 3 tendons; therefore, 253 repairs were used in the final analysis. There were 32 experimental groups with n=7 to 9 repairs/group (Appendix, Table A1).

Tendon repairs

Fresh-frozen cadaver arms were obtained through the Washington University in St. Louis School of Medicine’s Body Donor Program. A random number generator was used to assign the order and type of repair for each specimen. A postgraduate orthopaedic surgeon research fellow (EN) performed about one quarter of the repairs; a 3^rd year orthopaedic surgery resident (GN) performed the rest. The repair order was randomized prior to initiation of the study; repair techniques were proportionally distributed between the two surgeons. A primary suture repair utilizing a 4-0 or 3-0 caliber double-stranded suture was performed (Fig. 1). Either a 4-strand modified locking Kessler technique or an 8-strand Winters-Gelberman technique was employed.⁴ Core suture purchase was either 0.75 or 1.2 cm from the cut tendon edge. After the core suture was tensioned so that the tendon edges were apposed, the peripheral suture was placed (Fig. 1). A 6-0 or 5-0 caliber suture was employed in a superficial or a deep configuration. Detailed methods can be found in the Appendix.

Figure 1.

A primary suture repair of caliber 4-0 or 3-0 was performed on cadaver flexor tendons. Either a 4-strand modified locking Kessler technique or an 8-strand Winters-Gelberman technique was employed. Core suture purchase (indicated by ‘x’ on the left schematics) was either 0.75 or 1.2 cm from the cut tendon edge. After the core suture was tensioned so that tendon edges were apposed, the peripheral suture was placed. A 6-0 or 5-0 caliber suture was employed in a superficial or a deep configuration. The superficial suture was placed into the tendon <1 mm from the cut end (indicated by ‘x’ on the right schematics) and <1 mm deep (indicated by ‘y’ on the right schematics) within the tendon substance. The deep variant was placed 2 mm from the cut tendon edge and ~ 2 mm deep within the tendon substance.

Biomechanical testing

Repairs were tested biomechanically using methods described previously (Appendix).^15-17 Briefly, the proximal tendon was gripped with a triangle-toothed grip, and the bone of the distal phalanx was gripped in a custom grip (Fig. 2). Tendons were then preconditioned and tested in tension until failure. Failure mode (e.g., suture pullout or suture breakage) was recorded. The extension at 20N (mm) (a typical load seen during passive motion rehabilitation^{18, 19}), load at 2 mm gap (N), maximum load (N), extension at maximum load (mm), stiffness (N/mm) (the slope of the linear portion of the load-deformation curve), and modified toughness (N*strain) (the area under the load-strain curve) were determined using a custom-written code in MATLAB (Natick, MA). Load at a gap of 2 mm (a threshold level that leads to decreased repair strength and increased adhesions^{3, 20}) was calculated using tracking regions nearest to the repair site.

Figure 2.

(Left) As seen in the white dotted box, a random texture was applied to the tendon surface using Verhoeff‘s stain. The proximal tendon was gripped with sandpaper and a triangle toothed grip. Distally, the bone of the distal phalanx was gripped with 2 side screws and a pin in a custom grip. (Right) Load-deformation and load-strain curves are shown for a typical test. Biomechanical outcomes are indicated on the curves.

Statistical methods

Data analysis was generated using SAS (SAS v9.2 Institute Inc., Cary, NC). Outcomes across the 5 parameters (core suture, strands, core purchase, peripheral suture, and peripheral purchase) were compared using Generalized Estimating Equations (GEEs). Subgroup effects for the number of core suture strands were examined because (a) the determination of the number of strands may be subject to surgeon bias and (b) an 8-strand repair may not always be possible. Ancillary GEE models were performed to test the interaction effect that was considered clinically relevant: core suture caliber and number of core suture strands. Stepwise multiple linear regressions were performed to examine factors associated with select outcomes (extension at 20N, load at 2mm gap, and stiffness). Detailed methods can be found in the Appendix.

RESULTS

Demographic information

Data consisted of 47 cadavers (21 were female) with average age of 79 ± 12 yrs. Detailed specimen characteristics can be found in the Appendix (Table A2). The characteristics were not significantly different between the two surgeons.

The effect of surgical factors

The factors with the most widespread effects on the repair were the number of core suture strands and the peripheral suture purchase (Table 1). Factors that affected the fewest number of biomechanical properties were core purchase and peripheral suture caliber. Core suture caliber significantly affected 3 of the biomechanical properties studied.

Table 1.

Effect of surgical factors on the biomechanical properties of the effects. Table of pvalues from GEE model (values < 0.05 are highlighted in green).

	Ext. at 20N	Load at 2mm	Max. Load	Ext. Max. Load	Stiffness	Toughness
Core Suture Caliber	0.078	0.44	0.0012	<.0001	0.55	<.0001
Core Suture Strands	0.0018	<.0001	<.0001	0.0003	<.0001	<.0001
Core Suture Purchase	0.58	0.55	0.0046	0.15	0.60	<.0001
Periph Suture Caliber	0.40	0.66	0.57	0.0073	0.16	0.0030
Periph Suture Purchase	0.0020	<.0001	<.0001	0.0041	0.0009	<.0001

Core suture caliber: Moving from 4-0 to 3-0 caliber Supramid suture produced significant decreases in extension at maximum load and significant increases in maximum load and toughness (Table 1). Core suture strands: Increasing from 4 to 8 strands resulted in significant decreases in extension at 20 N and extension at maximum load (Figs. 3 and 4, Table 1). Significant increases were seen in load at 2 mm of gap formation, maximum load, stiffness, and toughness. Core suture purchase: Changing core suture purchase from 0.75 to 1.2 cm produced significant increases in maximum load to failure and toughness (Table 1). Peripheral suture caliber: Increasing peripheral suture caliber produced a significant decrease in extension at maximum load and a significant increase in toughness (Table 1). Peripheral suture purchase: Altering peripheral suture purchase from superficial to 2 mm resulted in significant decreases in extension at 20 N and extension at maximum load (Figs. 3 and 4, Table 1). Significant increases were seen in load at 2 mm of gap formation, maximum load, stiffness, and toughness.

Figure 3.

Biomechanical properties by core suture. Load at 2 mm gap, maximum load, and toughness results are shown for core purchase of 0.75cm. Results are grouped by core suture caliber and comparisons are shown for the two most important factors: core suture strands and peripheral suture purchase. Increasing the number of strands and increasing the peripheral suture purchase led to increases in all three outcome measures.

Figure 4.

Biomechanical properties by core purchase. Load at 2 mm gap, maximum load, and toughness results are shown for core suture caliber of 4-0. Results are grouped by core purchase and comparisons are shown for the two most important factors: core suture strands and peripheral suture purchase. Increasing the number of strands and increasing the peripheral suture purchase led to increases in all three outcome measures. Note the striking effect of peripheral suture purchase on load at 2 mm.

To determine whether these conclusions were valid when an 8-strand repair may not possible (e.g., when repairing a very small tendon), a separate model was run for each strand group. When the choice of strands was removed, the above conclusions remained unchanged (Appendix, Table A3).

When examining the interaction between core suture caliber and the number of core suture strands, only the outcome measure toughness revealed a significant interaction (Appendix, Table A4). There was no significant interaction effect between these two factors for the other 5 biomechanical outcomes.

The only observed failure modes were suture breakage and suture pullout. Suture knot unravelling, as reported in a recent study²¹, was not observed. Core suture caliber and the number of strands significantly affected failure mode (Fig. 5, Appendix, Table A5). Increasing the core suture caliber shifted the majority failure mode from breakage toward pullout. Similarly, increasing the number of strands shifted the mode from breakage to pullout. Core suture purchase, peripheral suture caliber, and peripheral suture purchase did not affect failure mode.

Figure 5.

Failure mode. Core suture caliber and the number of strands significantly affected the failure mode. Increasing the core suture caliber shifted the failure mode from suture breakage to suture pullout. Increasing the number of strands shifted the failure mode from suture breakage to suture pullout.

The effect of tendon cross-sectional area

Tendon CSA was significantly higher in males (Appendix, Table A2). CSA was not associated with core suture caliber, number of core strands, core purchase, or peripheral suture caliber (i.e., the random assignment resulted in no difference in CSA among groups with different factors) (Appendix, Table A2). There was a small, but significant association between CSA and the peripheral suture purchase (p = 0.02), suggesting that, despite the randomization process, tendons with larger CSAs were more likely to be assigned to repairs using a peripheral suture purchase of 2 mm. However, the mean CSA of repairs receiving a deep peripheral suture was 11.6 ± 2.4 and that of the superficial peripheral suture group was 11.0 ± 2.2. Though a significant difference, it is unlikely that this small difference exerts a clinically meaningful effect on our results.

The relevance of CSA to the biomechanical properties of the repair was examined for select outcomes (i.e., extension at 20N, load at 2 mm, and stiffness) where CSA was a significant factor (Appendix, Table A3). When the analysis was performed by strand, CSA was no longer relevant for extension at 20N and load at 2 mm. However, regardless of the number of strands, CSA remained associated with stiffness. Stepwise multiple linear regression models of the 5 factors, age, and CSA associated with the 3 biomechanical outcomes are described in Appendix, Table A6. There was a weak, but significant correlation between CSA, the number of core suture strands, and the peripheral suture purchase.

The effect of surgeon

A significant surgeon effect was found for extension at 20N, load at 2mm gap, maximum load, and toughness (Appendix, Table A7). There was no effect of surgeon for stiffness or extension at maximum load.

DISCUSSION

We described the biomechanical effects and interactions among 5 variables that surgeons consider when performing a flexor tendon repair. Previous ex vivo and in vivo studies described the individual effects of these variables. Dionoupolos et al. showed that, in addition to increased maximum load, 8-strand repairs resisted gap formation more effectively than did 4-strand repairs.²² Boyer et al. used an in vivo canine model to demonstrate that flexor tendons repaired with an 8-strand core suture had a significantly stronger load to failure during the first 3 wks in vivo when compared to 4-strand repairs.²³ In an ex vivo analysis, Taras et al. reported that increasing core suture caliber from 5-0 to 4-0 and from 4-0 to 3-0 in 2-strand repairs produced increases in mean tensile strength of 67% and 52%, respectively.⁵ Other ex vivo studies showed that increasing core suture purchase from 0.3 to 1.2 cm in 2- and 4-strand repairs improves resistance to gap formation, stiffness, and load to failure.^{7, 13, 24} In both an experimental and computational ex vivo study, core suture repair augmented with a deep peripheral suture displayed an 80% increase in strength compared to a superficial suture alone.⁶ Our study demonstrated that the factors producing the most widespread effects on biomechanical properties were the number of core suture strands and the peripheral suture purchase. Therefore, when repairing flexor tendon injuries to the middle three digits, surgeons should make all possible efforts to perform a multi-strand core suture repair including greater than 4 strands with a peripheral suture purchase of 2 mm. We found that core suture purchase and peripheral suture caliber had little influence on biomechanical properties. Relative to the effects of the number of core suture strands and peripheral suture purchase, it is of little consequence if the core suture is placed 0.75 or 1.2 cm from the repair site. There is also little effect of 5-0 compared to 6-0 caliber suture for the peripheral suture. The choice of core caliber marginally affected the biomechanical properties of the repair. When possible, 3-0 caliber suture should be chosen over 4-0 caliber suture, provided that future in vivo studies continue to demonstrate the safety of this method vis-à-vis tendon gliding properties and adhesion formation.

Our investigation of standard 4- and 8-strand flexor tendon repair techniques confirmed previously accepted observations. Increasing the number of core suture strands crossing the repair site increases the load bearing capacity of the repair. 8-strand repairs withstand gap formation more effectively than 4-strand repairs. Increasing core suture caliber increases maximum load to failure and shifts the failure mode from suture breakage to pullout. Proper depth of the peripheral suture is critical to preventing gap formation. These results, demonstrated in prior reports as isolated variables, were confirmed in our combinatory study design.

In vivo investigations revealed that gap formation is a major contributor to repair failure.²⁰ We determined repair-site extension at 20 N force and force at 2 mm gap as two measures related to gap resistance. The two technical factors significantly affecting extension at 20 N were the number of core suture strands and depth of peripheral purchase. Of the 5 repair techniques (out of the 32 tested) with the lowest mean extension at 20 N, four were 8-strand repairs and all employed 2 mm of peripheral suture purchase. A similar pattern was seen for strain at 20N. In examining the 5 techniques with the greatest load at 2 mm of gap formation, all five employed 2 mm of peripheral suture purchase. Four of these were 8-strand repairs. These observations are not surprising considering that core strand number and peripheral suture purchase were the most important factors affecting biomechanics based on our overall analysis. In particular, the results highlight the importance of the peripheral suture and its placement for repair-site resistance to gap formation.

Surprisingly, core suture purchase was one of the least important variables. This may be due to the fact that we investigated two relatively large values, 0.75 and 1.2 cm. Prior studies tested a range from 0.3 to 1.2 cm.^{7, 13, 24} Tang et al. showed that resistance to gap formation and ultimate strength increased significantly as suture purchase increased from 0.4 to 1.2 cm. However, strength was not different when comparing purchases of 0.7 to 1.2 cm.⁷ Therefore, beyond a 0.7 cm threshold, relatively little benefit is seen with increasing levels of purchase.

The failure mode gives an indication of the weakest link in the repair construct. Possible outcomes in our study included: (1) rupture of the suture material (indicating that the suture material is weaker than the suture grasping strength), (2) suture pullout (indicating that the suture grasping strength is weaker than the suture material), (3) tendon rupture away from the repair site (indicating that the tendon strength is weaker than the repair strength), (4) enthesis rupture (indicating that the tendon-to-bone attachment strength is weaker than the repair strength), and (5) tendon grip rupture or slippage (indicating that the gripping strength of the testing fixture was weaker than that of the repair). Since there were multiple core suture strands, a combined failure mode of suture rupture and pullout was also possible. Ruptures away from the repair site (options 3, 4, or 5) were not noted in any sample. Increasing the caliber of the core suture resulted in a shift from suture rupture to suture pullout and a moderate improvement in mechanical properties. This shift in failure mode suggests that the weak link in many repairs using smaller core suture caliber is the suture strength. Increasing the caliber of the core suture overcame this weak link, and the strength of the repair then depended on the grasping strength of the suture configuration.

Although the optimal clinical scenario would result in an 8-stranded core suture repair with a high-caliber core suture strand combined with a deep epitendinous suture, there are scenarios where this may not be possible. Given variations in tendon size²⁵, surgeons may be unable to perform a complex repair in all cases. For example, repair would be difficult when the flexor digitorum profundus tendon is transected deep to the A4 pulley in a finger with small caliber tendons. Also, distal flexor digitorum profundus lacerations within 1 cm of their insertions or transections of the flexor pollicis longus tendon within the substance of the thenar eminence can be difficult due to the constraints associated with the available space for applying various components of the repair. In contrast, a metacarpophalangeal joint level transection of both flexor digitorum superficialis and flexor digitorum profundus tendons of the middle finger provides optimal flexibility; both tendons have large cross sectional areas and sufficient space exits proximal to the A2 pulley for maximizing the repair’s biomechanical properties without anatomic constraints. These examples highlight the need for a combinatorial understanding of the biomechanical consequences for each repair technique variable.

Limitations must be consider in our study. First, this study was limited to repair-site biomechanics in cadavers and therefore represents “time zero” properties. Prior in vivo studies showed that each of the repair components studied here is well tolerated. However, not all combinations used in our study have been studied in vivo. Second, we employed 2 surgeons. While they worked together to standardize methods, there was a significant effect of surgeon on the properties of the repairs. Although the absolute values were different when comparing surgeons, the trends among experimental groups for each surgeon were the same. The differences between the surgeons may have been due to experience or variations in surgical technique. Both employed identical suture materials, and the randomization resulted in a proportional distribution of groups. The consistency in trends strengthens the conclusions that can be drawn from our study. Despite inter-surgeon variability, strong recommendations can be made regarding the factors studied with broad relevance to the community, which includes surgeons with varying levels of experience and skill.

Third, the gliding properties of the repairs were not examined. Function of a digit after flexor tendon repair depends on both repair integrity and gliding of the tendon within its sheath. Repair methods that increase friction between the surface of the tendon and its sheath may fail due to adhesion formation.^26-29 However, permutations of repair methods tested in our study are expected to have similar gliding characteristics.^{24, 30, 31} Moriya, et al. showed no difference between work of flexion for 2- vs. 4-strand techniques.³¹ Our repairs were performed with a looped suture employed in a locking Kessler configuration similar to Moriya’s modified Kessler. Using a looped suture in this fashion produces a repair with twice as many strands crossing the repair site, but achieved with the same number of needle passes (i.e., 4- and 8-strand techniques in our study had equivalent needle passes to 2- and 4-strand techniques in the published study). Based on the results comparing gliding in 2- and 4-strand repairs, we would expect no differences in gliding between 4- and 8-strand repairs. Alavania et al found no difference between 3-0 and 4-0 polyester 4-strand core suture repairs coupled with a running epitenon stitch.³⁰ While the core suture contributed to increased work of flexion if the caliber was > 3-0, there was no difference found when comparing 3-0 to 4-0. As our study also used core suture calibers of 3-0 and 4-0, we believe that there would be no difference in gliding properties for these two sutures. Lee et al established that work of flexion decreases with increasing core suture purchase from 3 to 10 mm, but that there was no significantly difference between 7 and 10 mm.²⁴ Based on these results, the gliding properties for 7.5 and 12 mm purchase, as performed in our study, would be similar.

Fourth, though a large number of repairs were tested, they were divided among 32 experimental groups, thus limiting the number of samples to between 7 and 9 per group. Only two variations of each of the 5 variables were tested. The objective of the current study was to identify the combinatorial effect of these 5 flexor tendon surgical technique factors on repair-site biomechanics. Previous reports narrowed the surgeon’s choices for each of the factors; however, the literature lacks information about which are most important, so we chose only two clinically significant options for each. As the numbers of repairs quickly rises as options and/or factors are added, certain limitations were accepted to allow a feasible study of this scope. Furthermore, two statistical arguments support the sample size: (1) our historical data from cadaveric and in vivo canine studies indicates that N=8 provides sufficient power of 80% assuming typical coefficients of variation and effect sizes for flexor tendon biomechanics^{8, 15, 25, 32, 33}; (2) although each group contained only 8 samples, the statistical design and analysis incorporates much larger numbers when evaluating differences between factors. For example, the sample size when asking the question of whether 8-strand repairs are different from 4-strand is effectively 128. At the next level of comparison (e.g., 8-strand with 3-0 caliber vs. 8-strand with 4-0 caliber vs. 4-strand with 3-0 caliber vs. 4-strand with 4-0 caliber), the sample size is effectively 64.

Fifth, a running epitenon (i.e., peripheral) suture technique was chosen. Several other techniques are employed by surgeons; the factors peripheral suture purchase and peripheral suture caliber may be more important to biomechanical properties of the repair with these techniques compared to running suture technique. However, to maintain a reasonable study design, we chose to investigate the most widely-accepted techniques based on our review. Future studies will be required to compare additional technique permutations.

To enhance initial repair site strength so that early rehabilitation can be undertaken (maximizing intrasynovial gliding and minimizing adhesion formation), we recommend that surgeons continue to pursue multi-strand repairs, as 8 strands crossing the repair offered substantially better biomechanical properties than did 4 strands. The addition of a peripheral suture with 2 mm purchase is also recommended to prevent gap formation and augment repair-site strength. Finally, since core suture caliber exerted significant effects on many biomechanical properties, the use of 3-0 suture could further enhance the properties of the repair.