 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1999 The Journal of Bone and Joint Surgery, Inc. The Effect of Gap Formation at the Repair Site on the Strength and Excursion of Intrasynovial Flexor Tendons. AN EXPERIMENTAL STUDY ON THE EARLY STAGES OF TENDON-HEALING IN DOGS* , ST. LOUIS, MISSOURIInvestigation performed at the Department of Orthopaedic Surgery, Washington University School of Medicine, St. Louis
Background: Elongation (gap formation) at the repair site has been associated with the formation of adhesions and a poor functional outcome after repair of flexor tendons. Our objectives were to evaluate the prevalence of gap formation in a clinically relevant canine model and to assess the effect of gap size on the range of motion of the digits and the mechanical properties of the tendons. Methods: We performed operative repairs after sharp transection of sixty-four flexor tendons in thirty-two adult dogs. Rehabilitation with passive motion was performed daily until the dogs were killed at ten, twenty-one, or forty-two days postoperatively. Eight tendons ruptured in vivo. In the fifty-six intact specimens, the change in the angles of the proximal and distal interphalangeal joints and the linear excursion of the flexor tendon were measured as a 1.5-newton force was applied to the tendon. The gap at the repair site was then measured, and the isolated tendons were tested to failure in tension. Results: Twenty-nine tendons had a gap of less than one millimeter, twelve had a gap of one to three millimeters, and fifteen had a gap of more than three millimeters. Neither the time after the repair nor the size of the gap was found to have a significant effect on motion parameters (p > 0.05); however, the ultimate force, repair-site rigidity, and repair-site strain at twenty newtons were significantly affected by these parameters (p < 0.05). Testing of the tendons with a gap of three millimeters or less revealed that, compared with the ten-day specimens, the forty-two-day specimens failed at a significantly (90 percent) higher force (p < 0.01) and had a significantly (320 percent) increased rigidity (p < 0.01) and a significantly (60 percent) decreased strain at twenty newtons (p < 0.05). In contrast, the tensile properties of the tendons that had a gap of more than three millimeters did not change significantly with time. Conclusions: Our data indicate that, in a dog model involving sharp transection followed by repair, a gap at the repair site of more than three millimeters does not increase the prevalence of adhesions or impair the range of motion but does prevent the accrual of strength and stiffness that normally occurs with time. Clinical Relevance: Tendons that have a large gap are at increased risk for rupture during early rehabilitation, and this risk of rupture does not decrease in the first six weeks.
Experimental and clinical efforts to improve functional recovery after repair of digital flexor tendons have included rehabilitation techniques that cause greater magnitudes of excursion of the intrasynovial tendons and higher levels of force on the tendons in vivo7,12,19,31,35. Clinical observations that elongation at the repair site occurs when these accelerated rehabilitation methods are used and that there is an inverse relationship between gap formation and clinical results indicate that early deformation is an important variable with regard to the recovery of tendon function10,13,14,27. Gaps as small as one to two millimeters at the repair site have been associated with poor functional performance and an increased need for a secondary tendon procedure for the lysis of peritendinous adhesions13,14,27. Recently, several investigators have attempted to define the tolerance limits for the application of increased in vivo force after tendon repair16,17,26. We do not believe that the extent to which excessive force leads to elongation at the repair site when early controlled passive and active mobilization methods are used or the effect of increased elongation on the strength of the repair has been evaluated with use of a clinically relevant animal model. Our objectives were to determine the prevalence of gap formation after repair of flexor tendons in dogs in which multistrand suture methods and early controlled motion had been applied and to quantify the effects of elongation at the repair site on digital motion and tendon strength. We hypothesized that an increase in the length of the gap would have a significant effect on the functional characteristics and the structural properties of repaired tendons.
We performed sixty-four tendon-repair procedures on the right forelimbs of thirty-two adult mongrel dogs (body mass, twenty to thirty kilograms). All of the operations, postoperative care, and rehabilitation were approved by the animal studies committee and were performed in a licensed animal-care facility. The animals were anesthetized with an initial intravenous dose of thiopental sodium (0.5 milliliter per kilogram of body weight), supplemented with intermittent injections of atropine (0.5 milliliter) and acepromazine (0.2 milliliter). The animals were intubated, and anesthesia was maintained with 1 percent halothane. The paws were shaved, washed with Betadine (povidone-iodine), and exsanguinated. The procedure was performed under tourniquet control. Midlateral incisions were used to expose the digital sheaths of the second and fifth digits in the region between the annular pulleys proximal and distal to the proximal interphalangeal joint. The sheath was entered, and the flexor digitorum profundus tendon was incised transversely. A four-strand repair was performed with double-stranded 4-0 braided caprolactam suture (Supramid; S. Jackson, Alexandria, Virginia), supplemented with a 6-0 nylon running epitendinous suture to invaginate the tendon ends36. The sheath was not repaired. Skin closure was performed with a running suture of 4-0 nylon. After the operation, the forelimb was placed in a fiberglass shoulder spica cast with the elbow at 90 degrees and the wrist at 70 degrees of flexion. The forelimb was immobilized at all times, except during two five-minute rehabilitation sessions that were performed five days a week, starting on the first postoperative day. A volar window in the cast was opened, and the digits were passively flexed and extended to the limit of the dorsal extension block. The wrist was maintained in the flexed position. Twelve animals were killed at ten days after the procedure; twelve, at twenty-one days; and eight, at forty-two days. Left and right forelimbs were disarticulated at the elbow, and they were stored at -20 degrees Celsius until the gap at the repair site was measured and biomechanical testing was performed38. The limbs were thawed to 23 degrees Celsius, and the second and fifth digits were disarticulated at the metacarpophalangeal joints. The flexor digitorum profundus tendons were transected proximally at the musculotendinous junction. Each repair site was exposed through a midlateral incision, with care taken to not disrupt the intact pulley system or any adhesions between the tendon and the intrasynovial sheath. The extent of adhesion formation was evaluated and was graded as none, mild, or severe. Eight tendons from eight dogs were found to be ruptured; these included four in the ten-day group, one in the twenty-one-day group, and three in the forty-two-day group. Twenty tendons in the ten-day group, twenty-three in the twenty-one-day group, and thirteen in the forty-two-day group were still intact. The length of the gap between the cut ends of these tendons was measured with digital calipers (Max-Cal; Fowler, Newton, Massachusetts) (Fig. 1).
The range of motion of the digits was then assessed with a computerized motion-analysis system (PC-Reflex; Qualisys, Glastonbury, Connecticut). Both the treated and the contralateral, control digits were evaluated. We used a procedure that was similar to one reported previously37 but that had been adapted for use with the motion-analysis system. Two cameras with infrared light sources were placed one meter from the region of interest, with their axes 60 degrees apart. Before data acquisition, the system was calibrated with use of a three-dimensional frame with six reflective markers at known positions. Two pairs of reflective hemispherical markers (four millimeters in diameter) were then pinned to the middle and distal phalanges of each digit and were reinforced with cyanoacrylate glue (Prism Adhesive; Loctite, Rocky Hill, Connecticut). A square piece of reflective tape (four millimeters) was glued to the flexor tendon. A baseline position was established by clamping the proximal phalanx, in a vertical orientation, to a rigid table and suspending a 0.15-newton weight from the tendon stump. The positions of the five markers were sampled for five seconds by the two cameras, and the three-dimensional coordinates were calculated. A flexed position was established by increasing the weight to 1.5 newtons, and the positions of the markers were resampled. The change in the angles of the proximal and distal interphalangeal joints and the linear displacement of the flexor tendon from the baseline to the flexed position were calculated assuming rigid body behavior. We did not assume that the proximal and distal interphalangeal joints rotated in the same plane, and we therefore used a three-dimensional technique9 for calculating the angles rather than the commonly reported two-dimensional technique34. The length of the gap at the repair site was measured again in order to determine whether the range-of-motion testing had increased the size of the gap. The specimens were prepared for tensile testing by dissecting the flexor digitorum profundus tendon and the distal phalanx from the digit. Four reflective markers were glued at one-centimeter intervals onto the volar aspect of the tendon, centered about the repair site (Fig. 2). The distal phalanx was held rigidly in a custom fixture that was attached to the actuator of a materials testing machine (model 8500R; Instron, Canton, Massachusetts). The proximal tendon stump was held stationary in a dry-ice freeze-clamp23, and the distal phalanx was displaced axially at a rate of 0.5 percent of the specimen length per second until failure. Force and marker positions were sampled during the failure tests. From the marker data, we calculated elongation-per-unit-length (millimeters per millimeter) for three subregions: across the repair site (between the second and third markers), distal to the repair site (between the first and second markers), and proximal to the repair site (between the third and fourth markers). The unit lengths were considered to be the initial distances between adjacent markers. The elongations distal and proximal to the repair site were typically less than 25 percent of the elongation across the repair site, and thus we report only data based on the repair-site elongation. The outcome parameters were ultimate force (newtons), rigidity (newtons per [millimeter per millimeter]) across the repair site, strain (percent) at the repair site at twenty newtons of force32, and strain at the repair site at failure. Ultimate force was defined as the highest value of force that was attained before failure, rigidity was defined as the slope of the linear region of the force versus elongation-per-unit-length curve, and strain at the repair site was defined as 100 times elongation-per-unit-length in the region between the second and third markers.
Linear regression analysis (performed with Statview 4.5; Abacus Concepts, Berkeley, California) was used to determine the effect of gap length, and analysis of variance was used to determine the effect of time on the outcome parameters.
The length of the gap after range-of-motion testing (mean and standard deviation, 1.64 ± 1.93 millimeters) was not significantly different from that before testing (mean, 1.66 ± 1.69 millimeters) (p = 0.46, Wilcoxon's rank test), indicating that our range-of-motion protocol did not lengthen the gap. We used the values that were obtained after testing for subsequent analysis. Ultimate force (p < 0.001; r2 = 0.19) and repair-site rigidity (p = 0.02; r2 = 0.11) were negatively correlated with the length of the gap. We therefore assigned the specimens to one of three groups according to the length of the gap. Twenty-nine tendons had a gap of less than one millimeter, twelve had a gap of one to three millimeters, and fifteen had a gap of more than three millimeters. Paired t tests indicated a significant (10 percent) decrease in distal interphalangeal joint flexion in the treated digits compared with that in the controls (p = 0.01). The values for proximal interphalangeal joint flexion (p = 0.81) and tendon excursion (p = 0.35) in the treated digits were not significantly different from those in the controls. These findings indicate that the treated digits had nearly the same range of motion ex vivo as the control digits did. With two-way analysis of variance, we could not detect a significant effect of either the time after the repair or the length of the gap on proximal interphalangeal joint flexion (time, p = 0.15; length, p = 0.61), distal interphalangeal joint flexion (time, p = 0.32; length, p = 0.69), or tendon excursion (time, p = 0.81; length, p = 0.54) in the involved digits (Table I).
In contrast, the time after the repair and the length of the gap had a highly significant effect on ultimate force, repair-site rigidity, and repair-site strain at twenty newtons in the treated digits (p < 0.001) (Table II). Values for ultimate force (Fig. 3-A) and rigidity (Fig. 3-B) at forty-two days were significantly greater than values at ten and twenty-one days for the tendons with a gap of less than one millimeter (p < 0.001) and those with a gap of one to three millimeters (p < 0.01). However, the force and rigidity values did not increase for the tendons with a gap of more than three millimeters. Repair-site strain at twenty newtons (Fig. 3-C) changed in a manner that was consistent with the changes in ultimate force and rigidity (that is, strain at forty-two days was significantly less than that at ten days for the tendons with a gap of less than one millimeter [p < 0.05] and those with a gap of one to three millimeters [p < 0.01], but it was unchanged for those with a gap of more than three millimeters). Repair-site strain at failure was the only tensile parameter that was not significantly affected by the time after the repair (p = 0.15) or the length of the gap (p = 0.26).
We noted no adhesions in thirty-six (64 percent) of the fifty-six intact tendons, mild adhesions in sixteen (29 percent), and severe adhesions in four (7 percent). We detected no association between the size of the gap and the presence of adhesions (p = 0.83, chi-square test).
Repair-site deformation occurred regularly in the early stages after repair of the intrasynovial flexor digitorum profundus tendon in our canine model. With use of a multistrand suture method and early controlled motion, a prevalence of rupture of 13 percent (eight of sixty-four) and a prevalence of substantial repair-site elongation (more than one millimeter) of 48 percent (twenty-seven of fifty-six) were observed. The rates of rupture and of gap formation in the present study are similar to those in several clinical studies in which controlled passive and active rehabilitation methods were used4, 22, 27, 30, 33. The importance of repair-site elongation, insofar as the recovery of tendon strength and excursion are concerned, has become increasingly controversial, with some authors believing that gap formation is one of the major causes of adhesion formation after tendon repair3, 10, 13, 14, 27. Seradge27, in a prospective clinical trial comparing two methods of tendon suture, noted a direct association between the occurrence of gaps between the stumps of repaired tendons and the formation of adhesions necessitating tenolysis. Although tenolysis was not necessary when there was repair-site elongation of less than one millimeter, the procedure was needed in three of nine digits in which there was two to three millimeters of tendon elongation and in seven of seven digits in which there was more than three millimeters of elongation. Other authors have had similar findings, suggesting that there is a relationship between formation of a gap and poor clinical results, and some of these authors have reported that the tolerance limits for repair-site defects were very small10, 13, 14, 27. The results of other clinical and experimental studies, however, have indicated that intratendinous gaps may not be deleterious under certain conditions. Although Mason and Allen21 noted that increased length of the gap was directly associated with decreased load on the tendon in vivo, they found that controlled motion both improved the resistance of the repair site to elongation and enhanced the quality of the tissue between the tendon stumps. Similar conclusions were drawn by Ejeskär and Irstam3 in a clinical study of repairs of the flexor digitorum profundus tendon in which only elongations of more than ten millimeters were associated with poor functional results. Silfverskiold et al.28 corroborated these observations in a clinical study in which elongation was measured with intratendinous metal markers that had been placed in the stumps of the flexor digitorum profundus tendon at the time of repair. They reported a mean final repair-site elongation of 3.2 millimiters, with only a week association between elongation and poor clinical results in terms of the active range of motion. In individual cases, the authors found that gaps were poor predictors of clinical results and that defects as large as ten millimeters were compatible with good functional recovery. Recent observations regarding the relationship between gap formation and final active motion of the interphalangeal joints therefore suggest that the deleterious effects of repair-site gaps may be obviated by early motion3, 28-30. However, to our knowledge, this important hypothesis has not been evaluated in a well controlled experimental model in which applied force and tendon excursion were known16,17 and repair-site elongation was quantified by direct measurements. The functional results in our study, which show that gap length did not adversely affect either interphalangeal joint motion or tendon displacement, support the conclusion that intratendinous repair-site gaps are well tolerated when rehabilitation methods involving early motion are used. Our observation that there was no increase in the formation of adhesions in tendons that had a gap as large as seven millimeters is consistent with the clinical findings of Ejeskär and Irstam3 as well as those of Silfverskiold et al.29, 30, but it is in contrast to the findings of Ketchum et al.13, 14 and those of Seradge27. We believe that controlled passive motion may have inhibited the formation of adhesions in our model, even in the presence of relatively large gaps. Although early controlled motion may minimize the adverse effect of repair-site elongation on the functional properties of the tendon, its effect on structural properties is not as well understood. Our findings indicate that the effect of gap size on tensile properties is time-dependent and that large gaps markedly impair the healing process (Figs. 3-A, 3-B, and 3-C). The strength and stiffness of the tendons that had a gap of three millimeters or less increased significantly with time (p < 0.01), so that at forty-two days the ultimate force before failure had increased by an average of 90 percent and the rigidity had increased by 320 percent compared with the values at ten days. However, when we tested the tendons that had a gap of more than three millimeters, we detected no significant difference between the values for ultimate force and rigidity at twenty-one or forty-two days and those at ten days. These findings suggest that tendons that have a large gap may be at increased risk for rupture throughout the rehabilitation period. Clinical concerns are therefore greatest in situations in which the rehabilitation program incorporates higher levels of in vivo tendon force. Specifically, at twenty-one and forty-two days after repair, tendons with a gap of more than three millimeters had a small margin of safety compared with the upper limit of tendon force (approximately twenty-nine newtons) required for active, unopposed digital flexion as reported by Schuind et al.26. Tendons with a gap of more than three millimeters failed at an average force that exceeded the in vivo force by only 33 percent at twenty-one days and by 52 percent at forty-two days compared with 132 percent at twenty-one days and 248 percent at forty-two days for tendons with a gap of less than one millimeter. Although the strength values used in this comparison were based on canine tendons, Noguchi et al.23 demonstrated that the ultimate forces before failure of repaired canine and human flexor tendons are similar and, thus, these comparisons should apply well to humans. These findings support efforts to develop repair methods that are less likely to cause gap formation and efforts to devise innovative techniques for assessing tendon elongation during the early intervals after repair 1,15,18,24,25. Both ultrasonographic planimetry and radiographic imaging with retained metal markers may be useful to assist clinicians in identifying digits in which greater levels of elongation have occurred22,28,30. Our finding that large gaps inhibited the accrual of tendon strength with time is consistent with findings of studies on the healing response of ligaments in the rabbit. Loitz-Ramage et al.20 and Chimich et al.2 reported that medial collateral ligaments that had no gap after injury had better mechanical strength than ligaments that had a gap of four to eight millimeters. Those and other studies suggest that gaps increase the time necessary for dense regular connective tissue to undergo the changes in collagen fibril diameter6 and collagen crosslink density5 that are needed to achieve normal mechanical integrity. One limitation of our study is that the dissection that was necessary to measure the length of the gap may have inadvertently caused disruption of adhesions in some specimens, thereby increasing the measured range of motion. Thus, it is possible that dissection may have contributed to the lack of difference that we found with respect to range of motion. We believe, however, that the effect of the dissection technique on our findings was negligible, as care was taken not to disrupt any visible adhesions. We demonstrated that the length of the gap before the range-of-motion testing did not differ from that after the testing. Therefore, in future studies, the gap can be measured after range-of-motion testing, to eliminate the possibility of disrupting adhesions. A second limitation relates to the measurement of the length of the gap. A gap was easily identified by noting separation of the cut ends of the tendon with immature repair tissue filling the gap (Fig. 1). However, precise measurement of the gap was difficult because the cut ends were not always parallel and the length of the gap was typically not uniform. We estimate the precision of the gap measurement to be within 0.5 millimeter of the actual length on the basis of repeated trials. Even though errors in gap measurement may have led to improper classification of a few tendons, we do not believe that the conclusions of our study were affected in any way. A third limitation relates to the relevance of the canine model to humans. The anatomy of the canine flexor tendon apparatus is similar to that in humans, and the relationships between tendon excursion and joint rotation are also similar11. It has been previously reported that increased adhesions8 and decreased functional properties (joint rotation and tendon excursion)7 result after immobilization in dogs, indicating that immobilization impairs tendon-gliding. These findings were consistent with clinical experience and support the relevance of the canine model for investigating repairs of the flexor tendons. Analysis of variance indicated that motion parameters were not significantly affected by the time after the repair or the length of the gap. To address the issue of the statistical power of these comparisons, we calculated the minimum detectable differences for proximal interphalangeal joint flexion, distal interphalangeal joint flexion, and tendon displacement using a significance level of 0.05 and a power of 0.8. Expressed as percentages of the overall sample means, the minimum detectable difference was 50 percent for proximal interphalangeal joint flexion, 28 percent for distal interphalangeal joint flexion, and 21 percent for tendon displacement. With the exception of the value for proximal interphalangeal joint flexion, these differences were close to the 25 percent level that we used for our original calculation of sample size, indicating that our conclusions of no significance are based on adequate statistical power. This experimental study demonstrates that, after a period of ten to forty-two days of daily controlled passive motion, the prevalence of tendon rupture and gap formation at the repair site was substantial. These findings support the observations of some investigators that repair-site elongation is relatively common in the early period after repair10,13,14,27. Furthermore, the observation that adhesions did not form in tendons that had a large gap is consistent with clinical data indicating that controlled passive motion may inhibit the formation of adhesions even in the presence of a gap3,29,30. However, the observation that gap formation had a negative effect on the strength of the repair site, as indicated by the significant reductions in ultimate force that were observed at twenty-one and forty-two days, suggests that tendons that have a gap of more than three millimeters may be at greater risk for rupture than those that have a gap of three millimeters or less. These findings demonstrate that rehabilitation methods that require higher levels of in vivo load should be carried out judiciously, particularly in situations where repair-site elongation is a concern.
NOTE: The authors gratefully acknowledge the operative assistance of Dr. Jingfan Zhang, Dr. Hitoshi Hatanaka, and Dr. Haralambos Dinopoulos and the technical assistance of Sara Ettner, Tim Morris, and Clint Walker.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Funds were received in total or partial support of the research or clinical study presented in this article. The funding source was National Institutes of Health Grant AR33097.
  CiteULike  Connotea  Del.icio.us  Facebook  Technorati  Twitter What's this?
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |