Abstract
The current trend toward early active flexion after repair of the flexor tendons necessitates a stronger repair than that provided by a modified Kessler technique with use of 4-0 nylon suture. The purpose of the current study was to determine, with use of the Taguchi method of analysis, the strongest and most consistent repair of the flexor tendons.Flexor tendons were obtained from fresh-frozen hands of human cadavera. Eight flexor tendons initially were repaired with the modified Kessler technique with use of 4-0 nylon core suture and 6-0 nylon epitenon suture. A test matrix was used to analyze a total of twenty variables in sixty-four tests. These variables included eight techniques for core-suture repair, four types of core suture, two sizes of core suture, four techniques for suture of the epitenon, and two distances from the repair site for placement of the core suture. After each repair, the specimens were mounted in a servohydraulic mechanical testing machine for tension-testing to failure.The optimum combination of variables was determined, with the Taguchi method, to be an augmented Becker technique with use of 3-0 Mersilene core suture, placed 0.75 centimeter from the cut edge with volar epitenon suture. The four-strand, double modified Kessler technique provided the second strongest repair. Five tendons that had been repaired with use of the optimum combination then were tested and compared with tendons that had been repaired with the standard modified Kessler technique. With the optimum combination of variables, the strength of the repair improved from a mean (and standard deviation) of 17.2 ± 2.9 to 128 ± 5.6 newtons, and the stiffness improved from a mean of 4.6 to 16.2 newtons per millimeter.
The technique of repair of the flexor tendons continues to evolve. Despite years of research by many investigators, the optimum method of suture remains elusive. As the understanding of healing and the biomechanics of the tendons advances, new goals are set for ultimate motion. Historically, treatment involved delayed primary tendon-grafting. Currently, primary repair is performed, followed by a postoperative protocol of active extension and passive flexion13. On the basis of the premise that early active flexion decreases the formation of adhesions and improves the ultimate range of motion and function5, there is now a trend toward primary repair followed by early active flexion with the goal of improving the final result23. Toward that end, investigators have tried to determine the method that provides the strongest repair9,14,25,27. Because of the many variables that are involved, the various testing methods that are used, and the different focus of each study, a review of the literature does not reveal the optimum combination of factors for the repair technique.
The purpose of the current study was to identify the factors most responsible for variations in the outcomes and to determine the value of the factors that result in the most consistent outcomes. Thus, the first goal of our study was to determine the method of repair that provides greater strength combined with less variation in the outcomes. To accomplish this, we used an experimental design and analysis called the Taguchi method16.
There are many examples, in the medical field, in which an optimum method or process was not determined because of the number of variables and the inherent limitations in the material available for study. Therefore, the second goal of our study was to evaluate the usefulness of the Taguchi method for future medical applications.
*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. No funds were received in support of this study.
†Read at the Annual Meeting of the American Society for Surgery of the Hand, Denver, Colorado, September 13, 1997.
‡Denver Orthopedic Specialists, Lutheran Medical Center Campus, 3550 Lutheran Parkway West, Suite 201, Wheat Ridge, Colorado 80033.
§J. Vernon Luck, Sr., Orthopaedic Research Center, Los Angeles Orthopaedic Hospital, 2400 South Flower Street, Los Angeles, California 90007-2697.
#Division of Plastic and Reconstructive Surgery, Department of Orthopaedic Surgery, University of California, Los Angeles, 200 UCLA Medical Plaza, Suite 140, Los Angeles, California 90095-6907. Please address requests for reprints to Dr. Jones.
**Department of Orthopaedic Surgery, University of California, Los Angeles, 100 UCLA Medical Plaza, Suite 305, Los Angeles, California 90024-6970.
Material
Flexor tendons (the flexor pollicis longus; the flexor digitorum superficialis of the index, long, and ring fingers; and the flexor digitorum profundus) were obtained from seven fresh-frozen hands of human cadavera. Each tendon was divided transversely into two sections of equal length. Comparison of proximal and distal sections of tendons has shown no substantial differences in the diameter or stiffness3. A total of fourteen flexor tendon specimens from each hand were available for testing. The size of the tendon is not under the control of the surgeon and was considered extraneous (so-called noise) according to the Taguchi method16.
Static Testing
After each repair, the specimens were mounted in a servohydraulic mechanical testing machine (Instron, Canton, Massachusetts) for tension-testing to failure. A specially designed jig was used to grip the ends of the tendon so that slippage would be minimized. The gripper first was tested with the whole tendon and was noted to withstand tensile loads of more than 1000 newtons before slippage was visually identified. Elongation was produced at a constant speed of 0.33 millimeter per second. A preload of 2.0 newtons was applied before testing to remove slack in the system. Ultimate strength was determined on the basis of the peak load that was recorded.
Real-time recording, made with use of an x-y plot for displacement-load analysis, revealed that the initial portion of each curve was linear. Stiffness was determined according to the slope of this initial curve. The load was determined at a displacement of 2.0 millimeters, and the result was divided by 2.0, to yield a stiffness in newtons per millimeter. We recognize that cross-head displacement for stiffness provides only an estimate of the true gap. During testing, we noted that almost all of the displacement was from the relatively weak and compliant repair site. Increased stiffness therefore indicates less gap formation.
Variables
We studied five control factors: the technique for core-suture repair, the type of core suture, the size of the core suture, the technique for suture of the epitenon, and the distance of the suture from the site of the repair (Table I). Eight techniques for core-suture repair were studied (Fig. 1): (1) the modified Kessler-1 technique (a two-strand repair with use of a single knot within the site of the repair), (2) the modified Kessler-2 technique (an epitenon-first, two-strand repair with placement of the core-suture knot away from the site of the repair), (3) the Tsuge technique24 (a two-strand repair with use of a single knot outside the site of the repair), (4) the Kessler-Tajima technique (a two-strand repair with use of two knots within the site of the repair), (5) the Lee technique11 (a four-strand repair with use of two knots within the site of the repair), (6) the double modified Kessler technique (a four-strand repair with the modified Kessler technique repeated; a single suture with one knot outside the site of the repair is used to produce the four-strand core repair), (7) the augmented Becker technique1 (a four-strand repair with use of two knots outside the site of the repair), and (8) the Savage technique19 (a six-strand repair with use of a single knot within the site of the repair).
We studied four core-suture materials: monofilament nylon (Ethilon), braided nylon (Nurolon), polypropylene monofilament (Prolene), and braided polyester (Mersilene) (all manufactured by Ethicon, Somerville, New Jersey). The study also comprised two sizes of core suture (4-0 and 3-0), four techniques for epitenon suture (epitenon not sutured, epitenon sutured volarly only, simple suture, and crossed suture), and two suture distances (0.5 and 1.0 centimeter). (The suture distance refers to the interval between the site of the repair and the point of entry or exit of the core suture.)
All knots were of the three-throw square type. All repairs were performed by the same surgeon (G. S.). All modified Kessler techniques were done with use of the locking-loop modification described by Pennington.
Experimental Design: Taguchi Method16
The classic way to understand the effect of a variable is to keep all other factors constant while varying factors one at a time. This results in a full factorial study that both is time-consuming and requires extensive resources. In contrast, the Taguchi method allows many variables to be studied, with use of a minimum number of experimental runs, through a systematic choice of combinations of variables (Table II). Not all combinations are studied.
The set of sixteen experiments was repeated four times with the order randomized within each set of experiments. Thus, there was a total of sixty-four experiments, which provided a minimum of eight repetitions of any particular variable. For example, the modified Kessler-1 technique was performed eight times, the Ethilon core suture was used sixteen times, and the 4-0 suture was tested thirty-two times.
A full factorial study would have required 512 (4 x 4 x 8 x 2 x 2) separate combinations. Repeating each combination four times would have required 2048 experiments. With use of the Taguchi method, the five factors that were chosen, with different multiple levels for each factor, were investigated in a total of sixty-four experiments.
Initial and Confirmation Runs
To validate the results of the Taguchi method, the initial and optimum repair techniques were compared. Initially, a set of eight flexor tendons was repaired with use of a standard method: a modified Kessler technique with placement of 4-0 Ethilon core suture 0.5 centimeter from the lacerated ends and a simple epitenon stitch with 6-0 Ethilon. At the completion of the sixty-four experiments, the optimum combination was determined with use of the Taguchi method. A set of five flexor tendons with this predicted optimum combination then was tested. The resulting mean and standard deviation were compared with those obtained with the initial standard method to demonstrate the improvement that had been achieved.
Signal-to-Noise Ratio
With traditional statistical methods, means are used to compare the results and the standard deviations are employed to determine whether the difference between two groups is significant. With the Taguchi method, a different statistic, the signal-to-noise ratio, is used to compare results.
With the Taguchi method, the variables that are being studied are divided into those that can be controlled (control factors) and those that either cannot be controlled or are too expensive to control (noise factors). The greater the effect of the noise, the greater the inconsistency. The goal of the Taguchi method is not only to choose control factors that produce the desired result (for example, a stronger repair) but also to direct a process that is less sensitive to noise. Although noise cannot be eliminated, its effect can be minimized, producing a result that is both stronger and less variable. The signal-to-noise ratio takes into account the mean as well as the variation between results. Therefore, the signal-to-noise ratio analysis can be thought of as being two-dimensional, whereas regular analysis is only one-dimensional16.
When the experimental goal is to maximize the outcome variable, the signal-to-noise ratio, in decibels, is: -10 * log10([1/y12 + 1/y22 +…+ 1/yn2]/n), where y is the strength or stiffness of each repetition of the experiment (in this study, there were four repetitions). The signal-to-noise ratio is, in a sense, a combination of the mean and the variance. Mathematically, the ratio increases as the individual values become larger. Improved consistency or decreased variability between values also increases the ratio.
The signal-to-noise ratio for tensile strength and stiffness was calculated for each of the sixteen experimental combinations, with four repetitions (Table III).
We next calculated the signal-to-noise ratio for each component of the five control factors. The value of these ratios is simply the mean of the signal-to-noise ratios of all experiments containing that particular component of the control factor (Table IV). For example, the signal-to-noise ratio for tensile strength, considering the control factor of the type of suture and the component of Ethilon, is: 1/4 x (25.8 + 30.4 + 37.2 + 31.2) = 31.2 decibels (Table III, combinations 1, 2, 7, and 8).
Results with Use of the Taguchi Method16
Strength and stiffness were only indirectly related. Correlation of the signal-to-noise ratios of strength with those of stiffness yields a Pearson correlation coefficient of 0.51 (p = 0.022). This suggests that only about one-half of the variation in stiffness could be predicted by the strength. A larger ultimate strength therefore did not necessarily correlate with a larger stiffness. Ultimate strength often occurred at more than eight millimeters of displacement, whereas determination of stiffness was within the initial two millimeters of displacement. Diao et al. studied strength and stiffness after tendon repair. They found that tendons that had been repaired with deep peripheral suture failed catastrophically at four millimeters of displacement, whereas those that had been repaired with superficial peripheral suture failed gradually by a pattern of breaking and unwinding at sixteen millimeters of displacement. On the basis of the biology of tendon-healing, we predict that a strong repair that allows a gap of eight millimeters would not be successful. The performance of the repair therefore should be judged according to the strength as well as the stiffness.
A comparison of two signal-to-noise ratios is less intuitive than a comparison of two means. It is possible to correlate the decibel difference with the per cent difference with use of the formula: per cent change in value = (10X/20 - 1) x 100, where X = the change in the signal-to-noise ratio (in decibels). For example, a 1.0-decibel difference in strength between factors is equivalent to a 12 per cent difference in strength, and a 10.0-decibel difference is equivalent to a 215 per cent difference.
Technique for Core-Suture Repair
Of all of the variables that were studied, the core-suture technique had the greatest effect on both the strength of the repair (maximum change in signal-to-noise ratio, 10.5 decibels; maximum signal-to-noise ratio, 36.6 decibels) and the stiffness of the repair (maximum change in the signal-to-noise ratio, 5.9 decibels; maximum signal-to-noise ratio, 16.2 decibels) (Table IV).
Type of Core Suture
The signal-to-noise ratios for the different core-suture materials, with the exception of Nurolon, were similar (Table IV). Several repairs that were performed with Nurolon failed because the knot became untied. This did not occur with any of the other suture materials. Three-throw square knots were used throughout the study; four throws would have been unrealistic given the bulk of the knot. This led to occasionally low values as reflected in low signal-to-noise ratios.
Technique for Suture of the Epitenon
A moderate improvement was obtained with repairs that included the epitenon compared with repairs that did not involve the epitenon. A simple epitenon suture resulted in a signal-to-noise ratio of 31.2 decibels, which was 2.5 decibels better than that resulting from repairs that did not involve the epitenon (28.7 decibels). This meant that the repairs that included a simple epitenon stitch were approximately 102.5/20, or 33 per cent, stronger.
Size of Core Suture
The signal-to-noise ratio for strength with use of 3-0 Mersilene suture (31.0 decibels) was 1.7 decibels better than that with use of 4-0 Mersilene suture (29.3 decibels), and stiffness improved 1.8 decibels (Table IV).
Distance of Suture from Site of Repair
The distance of the core-suture placement from the cut edge had the least effect on strength and stiffness of any variable that was tested. Failures due to pull-out of the sutures, rather than to breakage, were more prevalent for the repairs that had been performed at a distance of 0.5 centimeter than they were for those that had been performed at a distance of 1.0 centimeter. Of the thirty-two repairs that had been performed at a distance of 1.0 centimeter, only four (13 per cent) failed due to pull-out. Of the thirty-two repairs that had been performed at a distance of 0.5 centimeter, eleven (34 per cent) failed due to pull-out. Pull-out of the core suture is, in a sense, a premature failure because the suture pulls out of the tendon before the suture itself fails.
The shorter suture distance yielded a higher signal-to-noise ratio for stiffness, an intuitively apparent result.
Optimum Combination of Factors
We found that the optimum combination of factors, based on both strength and stiffness, was an augmented Becker technique with use of 3-0 Mersilene core suture, placed 0.75 centimeter from the cut edge with volar epitenon suture. As anticipated, this exact combination had not actually been tested in the Taguchi array. A set of tendons was repaired with use of this optimum combination to confirm the results of the Taguchi analysis. Then the results of the initial standard combination were compared with those of the optimum combination, both predicted and tested (Table V). The results of the testing of the optimum combination confirmed that the predicted results were accurate.
The goal of repair of the flexor tendons is restoration of full motion of the digit. Historically, repair was followed by immobilization. Healing was thought to occur through an extrinsic process mediated by the flexor sheath18, and this was believed to require immobilization of the tendon with formation of adhesions. Final motion of the digit was, not surprisingly, limited. Lundborg and Rank as well as Gelberman and Manske5 found evidence that the flexor tendon has an intrinsic repair capability. This revision of the understanding of the healing process gave impetus to studies of protocols for postoperative motion.
Gelberman et al.6 found that tendon-healing in dogs could occur without formation of adhesions. In addition, mobilized tendons healed more rapidly and had greater ultimate strength than immobilized tendons. Ivy et al. confirmed the benefits of light active mobilization after a four-strand core-suture repair of lacerations in zone II of the flexor tendon. The rate of good and excellent results was nineteen (76 per cent) of twenty-five in the group that had been treated with light active mobilization compared with fourteen (56 per cent) of twenty-five in the group that had been treated (in a previous study) with passive motion after a two-strand core-suture repair.
Active flexion places increased demands on the repair before the healing tendon shares the load. Currently popular techniques do not result in a repair that is sufficiently strong to withstand the forces associated with mobilization. Many factors have been studied with the goal of creating a stronger repair. Most investigators have focused on different techniques for core suture7,14,25,27. Some have dealt with various materials9,25,26 for core suture and with different techniques15,26 for epitenon repair. It is difficult to compare these studies or to draw conclusions from them as they have involved different tendon models (dog, rabbit, chicken, and human), in vivo and in vitro experiments, different sizes of sutures, and different techniques for testing.
Some investigators have focused on gap formation1,21,22. Logic dictates that a gap at the site of a repair will fill with fibrous tissue, resulting in an inferior repair with regard to the strength, stiffness, and length of the tendon. Clinically, this will take the form of decreased total active motion and an increased rate of tears. Seradge, in a prospective clinical study with use of radiopaque markers, found a direct relationship between the amount of elongation at the site of the repair and the rate of secondary tenolysis. In contrast, Silfverskiold et al. found only a weak relationship between elongation and final motion of the interphalangeal joint.
Small et al., in a prospective clinical study, managed ninety-eight patients with early active flexion after a modified Kessler repair with use of 4-0 Ethilon or 4-0 Monofil core suture and 6-0 Prolene epitenon suture. They noted excellent or good results in ninety (77 per cent) of 117 digits, but eleven (9 per cent) had dehiscence. Cullen et al. reported a rupture in two (6 per cent) of thirty-one digits after use of a modified Kessler technique with 3-0 Tycron core suture and 6-0 Prolene epitenon suture.
These studies are encouraging because they provide evidence that motion is improved after early active flexion. However, the higher rate of rupture reflects the large load that is placed on the relatively weak repair.
In order to interpret the expected demands on a tendon repair, it is important to examine the forces that the tendon may generate. Schuind et al. measured in vivo forces during carpal tunnel release with use of a specially designed device and found maximum forces of 8.8, 34.3, and 117 newtons in association with passive motion, active unresisted motion, and tip pinch, respectively. Forces even higher than 34.3 newtons may be predicted to occur in association with early active motion in digits with edema.
Optimum Variables
As was mentioned earlier, the optimum combination of variables for repair was based both on strength and on stiffness (Table IV). The ideal choice for each variable, such as the technique or suture material, was the one that provided maximum strength as well as stiffness. When the maximum values did not coincide, a rationale for the choice of variable was provided.
The augmented Becker technique was designated as optimum as it yielded the highest signal-to-noise ratio for both strength and stiffness. Mersilene was designated as the optimum type of core suture as it performed nearly the best with regard to strength and clearly the best with regard to stiffness. Prolene was the second choice. A suture size of 3-0 yielded the highest values for both strength and stiffness. Volar epitenon suture was chosen as strength was considered to be more important than stiffness. The distance of the placement of the suture from the cut end of the tendon was the only continuous variable; a distance of 0.75 centimeter was chosen in order to optimize both strength and stiffness. A combination of variables other than the optimum one may be used, but a confirmation experiment should be done for that combination in order to avoid the effects of unexpected interactions.
The technique for the core suture was the most important variable (Table IV). As might have been predicted, the two-strand techniques resulted in the lowest values. The six-strand Savage technique yielded poorer results than had been expected, perhaps because of the inability of each suture across the repair to share the load evenly, leading to early failure that was due to overload on any one suture.
Schuind et al. estimated the maximum expected force during active flexion to be 34.3 newtons. Insertion of this value into the described formula gives a value of 30.7 decibels. This estimate represents the minimum signal-to-noise ratio required for tendon repair; this level was not achieved with any of the two-strand techniques (Table IV). A signal-to-noise ratio of more than 31.0 decibels was reached with both the double modified Kessler and the augmented Becker technique (Table IV). Either of these techniques should provide a repair that is strong enough to withstand early active flexion. The six-strand Savage technique also provided a ratio of more than 31.0 decibels, but we believe that method is too difficult and time-consuming to be widely accepted by surgeons.
As mentioned, when both strength and stiffness are taken into account, Mersilene is the first choice for the suture material and Prolene is the second choice (Table IV). Techniques that involve suturing into the tendon fibers, as opposed to across or down them, require more handling of the tendon. Braided suture generates more friction and tends to deform the tendon more than does monofilament suture. Technical considerations may therefore lead to the choice of Prolene. Ethilon would be the best choice if strength were the only criterion; however, it performed poorly with respect to stiffness and therefore is not ideal when overall performance is considered.
Increasing the size of the core suture improved both strength and stiffness substantially (Table IV) but not as much as would have been predicted on the basis of the material properties alone. Ethicon4 reported a suture strength of 13.0 newtons for 4-0 Mersilene and 18.0 newtons for 3-0 Mersilene; this represents a 38 per cent difference in strength between the two sizes. The Taguchi method demonstrated a 1.7-decibel improvement in strength with use of 3-0 core suture (signal-to-noise ratio, 31.0 decibels) compared with that obtained with 4-0 core suture (signal-to-noise ratio, 29.3 decibels); this represents a difference of approximately 22 per cent.
Flexor-tendon injuries outside of zone II10 would certainly be better repaired with the larger suture. However, because of the greater bulk of 3-0 sutures, additional studies are needed to evaluate their gliding characteristics before they can be recommended for repair of these lacerations. There is a clinical precedent for use of this size of suture to repair zone-II injuries. Cullen et al., in a study of thirty-one digits that were treated with early active motion after repair of a zone-II injury with use of 3-0 Tycron suture and a modified Kessler technique, reported that twenty-four digits (77 per cent) had an excellent or good result and two (6 per cent) had a rupture.
The addition of simple epitenon suture has been shown to have an impact on strength15,26. In the current study, volar epitenon suture was chosen to test the clinical situation in which suturing of only the volar aspect of the epitenon is technically feasible. Simple repair was chosen instead of volar repair because the mechanical performance was similar and the biology of tendon-healing suggests that unexposed tendon leads to less scarring. Simple epitenon repair was 2.5 decibels, or approximately 33 per cent, stronger than repair that did not involve the epitenon. Wade et al. found that the addition of a peripheral 6-0 polypropylene stitch improved strength by 12.7 newtons, with an ultimate strength of 31.3 newtons (a difference of 41 per cent). The results of the current study are in agreement with the finding of Wade et al. that a peripheral stitch adds substantial strength to the repair.
When the modified Kessler-1 technique (core suture first) was compared with the modified Kessler-2 technique (epitenon suture first), the signal-to-noise ratio for strength was found to have improved from 26.1 to 27.6 decibels. This represents an expected improvement in strength of 101.5/20, or 19 per cent, as a result of suturing the epitenon first. This finding corresponds well with that of Papandrea et al., who reported a 22 per cent improvement in strength in association with an epitenon-first repair in twenty-six matched tendons of dogs. Those authors used 4-0 braided polyester core suture and 6-0 braided polyester epitenon suture.
Comparison of the Standard and Optimum Combinations
Taguchi analysis was used to identify a method that produced a stronger and less variable repair. The confirmation experiment was performed to determine whether the Taguchi method was accurate. Variation may be measured by the standard deviation, but the minimum result is also important. It is the occasional low value that is associated with the occasional failure.
The optimum method for repair improved strength to 128 ± 5.6 newtons compared with 17.2 ± 2.9 newtons with use of the standard method for repair. The low value of 24 per cent below the mean for the standard method improved to 5 per cent for the optimum method (Table V). Stress as high as 35.0 newtons can be anticipated at the site of a repair during unopposed active flexion; therefore, a repair that can resist 128 newtons of tension with minimum variation should give the surgeon enough confidence to begin a protocol for early postoperative active flexion.
The increase in stiffness from 4.6 newtons per millimeter with the standard method to 16.2 newtons per millimeter with the optimum method substantially decreases the gap under physiological load. With the standard combination, a load of 34.0 newtons would exceed the expected strength of the repair, but if the repair remained intact a gap of 7.4 millimeters is predicted. With the optimum combination, a maximum gap of 2.1 millimeters is expected at 34.0 newtons of load.
Taguchi Method
The Taguchi method differs from traditional statistical methods in its focus on identifying a solution that is both better and less variable; in the present study, this meant a stronger and less variable repair. The parameter used for optimization is the signal-to-noise ratio. In the formula for that ratio, a low value is more heavily penalized than a high value is rewarded because the low values are associated with failure. Not only does the signal-to-noise ratio help to reduce variability, it does so by identifying the control factors that lead to low values. The Taguchi method allows the factors that contribute the most to variation to be identified and facilitates determination of the settings or values that result in the least variability. However, it does not allow easy analysis of interactions between factors.
The Taguchi method has at least two limitations. First, it is not a pure statistical method. Statistics are used as a foundation, but a high degree of judgment is necessary for application. The test matrix (orthogonal array) allows testing of multiple variables, but only the most important interactions can be studied. Therefore, the investigator must have a good understanding of the process being studied in order to decide which factors to consider. Second, the Taguchi method does not provide innovation; it merely optimizes the combination of factors already known.
The augmented Becker technique, the optimum method for repair that was identified in the current study, may not be acceptable to many surgeons because, for example, of the time and effort that is required to perform it. The surgeon may decide that any four-strand repair technique is technically unappealing or is injurious to the tendon, or both. The signal-to-noise ratios demonstrated that two-strand techniques result in repairs that are simply too weak to allow for a reliable early active-motion protocol. As mentioned, both Small et al. and Cullen et al. used a two-strand repair technique followed by early active flexion in clinical studies. Small et al. reported that ninety (77 per cent) of 117 digits had an excellent or good result and that eleven (9 per cent) had dehiscence. Cullen et al. reported that twenty-four (77 per cent) of thirty-one digits had an excellent or good result and that two (6 per cent) had a rupture. The current study suggests that the Taguchi method may be useful in identifying the optimum combination for flexor-tendon repair, possibly leading to a better range of motion and a lower rate of rupture. Only a four-strand technique provides a repair that is strong enough to reliably allow immediate active flexion postoperatively.
Becker, H., and Davidoff, M.: Eliminating the gap in flexor tendon surgery. A new method of suture. Hand,9: 306-311, 1977.9306
1977
[PubMed]
Cullen, K. W.; Tolhurst, P.; Lang, D.; and Page, R. E.: Flexor tendon repair in zone 2 followed by controlled active mobilisation. J. Hand Surg.,14-B: 392-395, 1989.14-B392
1989
Diao, E.; Hariharan, J. S.; Soejima, O.; and Lotz, J. C.: Effect of peripheral suture depth on strength of tendon repairs. J. Hand Surg.,21A: 234-239, 1996.21A234
1996
Ethicon Knot Tying Manual, p. 39. Somerville, New Jersey, Ethicon, 1983.
Gelberman, R. H., and Manske, P. R.: Factors influencing flexor tendon adhesions. Hand Clin.,1: 35-42, 1985.135
1985
[PubMed]
Gelberman, R. H.; Vande Berg, J. S.; Lundborg, G. N.; and Akeson, W. H.: Flexor tendon healing and restoration of the gliding surface. An ultrastructural study in dogs. J. Bone and Joint Surg.,65-A: 70-80, Jan. 1983.65-A70
1983
Greenwald, D. P.; Hong, H. Z.; and May, J. W., Jr.: Mechanical analysis of tendon suture techniques. J. Hand Surg.,19A: 641-647, 1994.19A641
1994
Ivy, R. E.; Strickland, J. W.; Cannon, N. M.; and Gettle, K. H.: Early AROM program for zone II flexor tendon lacerations. Read at the Annual Meeting of the American Society for Surgery of the Hand, San Francisco, California, Sept. 14, 1995.
Ketchum, L. D.: Suture materials and suture techniques used in tendon repair. Hand Clin.,1: 43-53, 1985.143
1985
[PubMed]
Leddy, J. P.: Flexor tendons—acute injuries. In Operative Hand Surgery, edited by D. P. Green. Ed. 3, vol. 2, pp. 1823-1851. New York, Churchill Livingstone, 1993.
Lee, H.: Double loop locking suture: a technique of tendon repair for early active mobilization. Part I: Evolution of technique and experimental study. J. Hand Surg.,15A: 945-952, 1990.15A945
1990
Lundborg, G., and Rank, F.: Experimental intrinsic healing of flexor tendons based upon synovial fluid nutrition. J. Hand Surg.,3: 21-31, 1978.321
1978
Meals, R. A.: Current concepts review. Flexor tendon injuries. J. Bone and Joint Surg.,67-A: 817-821, June 1985.67-A817
1985
Noguchi, M.; Seiler, J. G., III; Gelberman, R. H.; Sofranko, R. A.; and Woo, S. L-Y.: In vitro biomechanical analysis of suture methods for flexor tendon repair. J. Orthop. Res.,11: 603-611, 1993.11603
1993
[PubMed]
Papandrea, R.; Seitz, W. H., Jr.; Shapiro, P.; and Borden, B.: Biomechanical and clinical evaluation of the epitenon-first technique of flexor tendon repair. J. Hand Surg.,20A: 261-266, 1995.20A261
1995
Peace, G. S.: Taguchi Methods: A Hands-on Approach to Quality Engineering, pp. 1-236, 292-312. Reading, Massachusetts, Addison-Wesley, 1993.
Pennington, D. G.: The locking loop tendon suture. Plast. and Reconstr. Surg.,63: 648-652, 1979.63648
1979
Potenza, A. D.: Critical evaluation of flexor-tendon healing and adhesion formation within artificial digital sheaths. An experimental study. J. Bone and Joint Surg.,45-A: 1217-1233, Sept. 1963.45-A1217
1963
Savage, R.: In vitro studies of a new method of flexor tendon repair. J. Hand Surg.,10-B: 135-141, 1985.10-B135
1985
Schuind, F.; Garcia-Elias, M.; Cooney, W. P., III; and An, K. N.: Flexor tendon forces: in vivo measurements. J. Hand Surg.,17A: 291-298, 1992.17A291
1992
Seradge, H.: Elongation of the repair configuration following flexor tendon repair. J. Hand Surg.,8: 182-185, 1983.8182
1983
Silfverskiold, K. L.; May, E. J.; and Tornvall, A. H.: Gap formation during controlled motion after flexor tendon repair in zone II: a prospective clinical study. J. Hand Surg.,17A: 539-546, 1992.17A539
1992
Small, J. O.; Brennan, M. D.; and Colville, J.: Early active mobilisation following flexor tendon repair in zone 2. J. Hand Surg.,14-B: 383-391, 1989.14-B383
1989
Tsuge, K.; Ikuta, Y.; and Matsuishi, Y.: Repair of flexor tendons by intratendinous tendon suture. J. Hand Surg.,2: 436-440, 1977.2436
1977
Urbaniak, J. R.; Cahill, J. D.; and Mortenson, R. A.: Tendon suturing methods: analysis of tensile strengths. In American Academy of Orthopaedic Surgeons: Symposium on Tendon Surgery in the Hand, pp. 70-80. St. Louis, C. V. Mosby, 1975.
Wade, P. J.; Muir, I. F.; and Hutcheon, L. L.: Primary flexor tendon repair: the mechanical limitations of the modified Kessler technique. J. Hand Surg.,11-B: 71-76, 1986.11-B71
1986
Wray, R. C., and Weeks, P. M.: Experimental comparison of technics of tendon repair. J. Hand Surg.,5: 144-148, 1980.5144
1980