- Email: [email protected]
Analysis of finger extensor mechanism strains
Analysis of Finger Extensor Mechanism Strains Patrick T. Hurlbut, MD, Burlington, VT, Brian D. Adams, MD, Iowa City, IA Strains in the extensor mechanism of the finger were measured in a cadaver model using Halleffect transducers. Several components of the mechanism were evaluated at different joint positions, with different intrinsic and extrinsic tendon loading conditions, and after creating a boutonniere deformity. Landsmeer's theory that predictable and obligatory interactions occur within the extensor mechanism during finger movement is strongly supported by our results. The concept of the Bunnell intrinsic-tightness test was confirmed. Results were consistent with clinical observations and current theories on the pathomechanics of claw and boutonniere deformities. Based on our experimental findings, we conclude that strain analysis is an effective method of evaluation of the extensor mechanism with potential for in vivo surgical applications. (J Hand Surg 1995;20A:832-840.)
The extensor mechanism of the finger is an elaborate system that integrates the forces of the lumbrical, interossei, and long extensor to produce precise functional movements. Although detailed descriptions of the functional anatomy have been reported, biomechanical support for current theories on extensor mechanics is limited. Previous studies have focused primarily on specific intricacies of the extensor mechanism. A strain analysis of the extensor mechanism was performed by Sarrafian et al.; however, the contribution of the intrinsics was not included. 1 Micks and Reswick developed a mathematical model based on the effective moment arms of the extensor components about the proximal interphalangeal (PIP) joint. 2 The model predicted strains in the central slip and lateral band but considered only PIP joint motion. Mulder and Landsmeer investigated extensor mechanism strains about the metacarpophalangeal (MP) From the University of Vermont, Depaltment of Orthopaedics and Rehabilitation, Burlington, VT and the University of Iowa, Department Orthopaedic Surgery, Iowa City, IA. Received for publication March 21, 1994; accepted in revised form March 17, 1995. 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. Reprint requests: Brian D. Adams, MD, Associate Professor, Division of Hand and Microsurgery, Department Orthopaedic Surgery, University of Iowa Hospitals & Clinics, Iowa City, IA 52242.
832
The Journal of Hand Surgery
joint associated specifically with the claw deformity. 3 In a study of geometric alignment and tensile properties of the extensor mechanism, Garcia-Elias et al. demonstrated significant variation among components, but deformities were not considered? Defining the interaction of forces within the extensor mechanism is important to the understanding of normal extensor function as well as the pathomechanics of many finger deformities. The goal of this study was to test contemporary theories on extensor mechanics of the normal and pathologic finger using a tissue strain model. Strain is a measure of tissue deformation resulting from an applied force. Although strain is an indirect indicator of force, it can be measured, unlike load, with minimal disturbance of tissues. To avoid the limitations of previous studies, an experiment was designed to evaluate the extensor mechanism in different composite finger positions and with different intrinsic and extrinsic tendon load conditions. During this experiment, we sought to assess the feasibility of using this method for in vivo evaluation of the extensor mechanism.
Materials and Methods Twelve fresh-frozen human cadaver long fingers amputated at the level of the proximal metacarpal were used. Specimens were thawed prior to testing and kept moist with normal saline. Dorsal skin and subcuta-
The Journal of Hand Surgery / Vol. 20A No. 5 September 1995
neous tissues were removed. A 1/8-inch steel rod was inserted into the medullary canal of the metacarpal and fixed with bone cement. The construct was mounted in a testing jig that allowed the MP and interphalangeal (IP) joints free range of motion. Fingers were tested in six different positions with composite angles of IP and MP joint flexion set at extension (0~ partial flexion (90~ midflexion (180~ full flexion (260~ intrinsicplus (60~ and hook (180 ~ (Fig. 1). To achieve these positions, monofilament nylon lines were attached to the tendons of the extensor digitorum communis (EDC), interossei, and lumbricals and routed over low friction pulleys to free weights. The proximal ends of the flexor digitorum superficialis and profundus tendons were clamped at their resting lengths in each of the six positions. Although an infinite number of tendon load combinations could potentially achieve a given finger position, previously reported electromyographic (EMG) data were used to determine muscle activity for each specific finger position?-8 Using this information, the following constraints were applied: (1) lumbrical activity is always present with IP joint extension,
regardless of MP joint position and (2) interossei have their greatest activity during IP joint extension with simultaneous MPjoint flexion. In addition, some interossei force is present in the fully flexed and extended positions as well as during finger flexion to provide smooth synchronous motion of all joints. However, intrinsic activity is absent in the claw d e f o r m i t y . 9 Since EMG data correlate poorly with muscle cell recruitment and do not predict tension magnitude,~~ the tension fractions of the muscles were used to specify the magnitudes of applied loads, u Three loading conditions of increasing weight were used. In the first loading condition, the loads used for the active muscles in a given finger position were set at 5% of their maximum predicted tensions (MPT), with some variations to satisfy the other experimental conditions (Fig. 1). The second and third loading conditions were determined by multiplying the percentage of MPT used in the first loading condition by 2 and 5 times, respectively. Thus, loads in the second condition were approximately 10% of MPT and those in the third condition were approximately 25% of MPT. The two
F~C
Finger P ~ i t t o n
c:::=::t~:::x::~
833
tl
Composite AngLe
Loed Conditi~'~
~PT
LOOd(e)
Ext4mlton
I
$
475
2
10
3
25
~g4PY
UI
L
Loadtg)
NEPT
Loed(g)
INPT
Z
250
2
~O
15
Loci(g) 150,
950
4
500
~
1,00
30
300
:X73
10
1250
10
1000
75
nO
O"
Portia( FLexion
:
r~ I
I
;-~,I
:
:
I~
I
C,I
i
i n u n u u l i I
90"
I;I~
IJ,
t*,I
I I , I
I
I'~-~I ,
I I , I
I[
I
I
,:,:,I
5
675
2
250
2
200
2
10
950
&
500
6
400
5 5
3
25
2375
10
1250
10
1000
5
Ftexion 180"
FuLL
1
5
~?~
2
25o
2
~
5
Ftexion 260"
2
10
950
4
500
4
40Q
5
3
25
z375
10
1~$0
10
100Q
5
Intrtnelr Ptul
Nook 180"
1
2.5
237.5
5
625
5
500
15
150
2
5
475
10
~250
10
3000
30
300
3
12.5
1~87.5
25
3125
25
25~
75
730
1
5
4?3
8
9
5
2
10
950
8
9
5
$
25
1187.5
8
9
5
Figure 1. Loading conditions for each corresponding finger position. %MPT, percentage of maximum predicted tension; EDC, extensor digitorum communis; RI, radial interosseous; UI, ulnar interosseous; L, lumbrical; Load(g), load in grams.
834 HurlbutandAdams/ Analysisof FingerExtensorMechanismStrains higher loads were applied for two purposes: first, to assure a response of the strain gauges to changes in load, and second, to determine if the ratio of strains among the extensor components varies with applied loads due to differences in the elastic properties among the components. All experimental conditions were satisfied when (1) EMG constraints were met, (2) the first loading condition held the finger in the desired position, and (3) finger position was maintained when weight was increased to the second and third loading conditions (Fig. 1). Strains measured under these conditions will reflect total force resulting from applied load and passive stretch. For example, in the intrinsicplus position there are high intrinsic forces but least passive stretch acting on the lateral bands. The opposite occurs in the hook position. Extensor mechanism strains were measured using Hall-effect transducers that were customized for this application (Micro Strain Inc., Burlington, VT). The gauges were attached to the extensor mechanism by short barbs on the ends of the device and 5-0 nylon suture (Fig. 2). Attachment was simple, quick, and required only 4 to 8 mm of tissue for each gauge. The
gauges did not cause tissue buckling, twisting, or other signs of interference with extensor function. The barbs and sutures held the gauges secure during testing and there was no grossly identifiable tissue damage after their removal. Gauges were mounted on the extensor mechanism after nominal loads (5 grams) were applied to the intrinsic and extrinsic tendons to remove observable slack from the tendons and lines. Baseline lengths measured at nominal loads were used for strain calculations. Regression analysis was performed to determine instances of significant specimen effect. Comparisons were derived using Fisher's least significant difference test and Bonferroni's multiple comparisons procedure to establish significant differences among conditions. Bonferroni's procedure was included to accommodate for the high number of comparisons. Least squares means were calculated to adjust for occasional missing data points. SAS software was used for statistical analysis with statistical significance taken at the 95% level (p < .05). Each specimen was tested using a combination of two of the following three protocols.
~
HII ~ Tmd~sduow Termlnll Two:Ion
d Exmnsor D - - m
//
Comrnunlm
//
Hall ~ TmnKl~[mr FIBdlIII ~ 1
\ Figure 2. Radial view of finger showing the attachment sites for three of the Hall-effect transducers. The ulnar lateral band implantation site (not illustrated) corresponds to a location similar to that shown for the radial lateral band. The fingers were mounted in a specially designed jig (bottom left).
The Journal of Hand Surgery / Vol. 20A No. 5 September 1995 835 Strain Distribution versus Finger Position In all 12 specimens, strains were measured at three sites on the intact extensor mechanism: the central band (CB) of the EDC tendon proximal to its junction with the oblique fibers from the lateral bands, the ulnar lateral band (ULB) proximal to its junction with the oblique fibers from the central band, and the terminal tendon (TT) distal to the triangular ligament (Fig. 2). Testing was done in all six finger positions using the second loading condition.
Second, the triangular and transverse retinacular ligaments were cut parallel with the fibers of the lateral bands. In the final step, the PIP joint was flexed to cause palmar subluxation of the lateral bands. Strains were measured after each step with the finger in full extension. After testing in full extension, the digit was tested in partial flexion, midflexion, full flexion, and the intrinsic-minus position.
Results Strain Distribution versus Finger Position
Strain Distribution versus Tendon Load In 6 of 12 specimens, strain measurements were made for all three loading conditions. In addition to the measurements made on the CB, ULB, and TT, strains were measured on the radial lateral band (RLB) proximal to its junction with the oblique fibers from the central slip (Fig. 2). Testing was done in all six finger positions. Claw and Boutonni?~re Deformities In the remaining six specimens, the full flexion position, hook position, and claw deformity were tested with no loads applied to the interossei and lumbrical tendons and 10% of MPT applied to the EDC tendon. Strains were measured at three locations (CB, ULB, and TT). A boutonniere deformity was then created using the sequence of cuts described by Zancolli. 9First, the central slip of the EDC tendon was cut near its attachment to the proximal phalanx.
Mean strain in the CB increased with increasing finger flexion, with 64% greater strain at full flexion. CB strain in midflexion, full flexion, and the hook position was significantly greater than in full extension (p < .05). Greatest CB strain occurred in the hook position and at full flexion. ULB strain decreased with increasing finger flexion, with 40% less strain in full flexion. Greatest ULB strain occurred at extension and in the hook position. Least ULB strain occurred in the intrinsic-plus position. ULB strain in full flexion and in the intrinsic-plus position was significantly less than in full extension (p < .05). TT strain increased with increasing finger flexion, with 124% greater strain in full flexion than in full extension. Greatest TT strain occurred in full flexion and in the hook position. Partial flexion, midflexion, full flexion, and the hook position caused significantly greater TT strain than occurred in extension and the intrinsic-plus position (p < .05) (Fig. 3).
6 5
% 4
r a n
Central Band
m~] Ulnar Lateral Band
[
I Terminal Tendon
Figure 3. Strains in the central band, terminal tendon, and ulnar lateral band as a function of finger position (n= 12). Error bars denote 1 SD.
836
Hurlbut and Adams / Analysis of Finger ExtensorMechanism Strains
Strain Distribution versus Tendon Loads
Increasing loads on the tendons did not alter the relative strain distribution for each finger position. Strains increased progressively in the CB, RLB, ULB, and TT with heavier load conditions. For example, Figure 4 demonstrates the changes that occur in the TT with increasing load. Claw and Boutonniere Deformities
Mean strains in the claw deformity, hook position, and full flexion position were evaluated for significant differences. CB strain was less in the claw deformity than in full flexion (p < .05) or the hook position (p = .07). ULB strain was greater in the hook position than in the claw deformity or in full flexion (p < .05); in fact, ULB strain was 10 times greater in the hook position than in full flexion (p < .05). No significant differences in TT strain occurred among the finger positions (Fig. 5). CB strain decreased with each step during creation of the boutonniere deformity (Fig. 6). ULB strain increased slightly after the central slip was cut and again increased slightly after the triangular and transverse retinacular ligaments were cut. ULB strain then decreased slightly after the lateral bands migrated palmarly. TT strain decreased slightly with the first cut but recovered after the lateral bands migrated palmarly. Strains in each complete boutonniere deformity were compared to those in the intact digit. CB strain in the boutonnibre deformity averaged 60%
less for all finger positions (p < .05) (Fig. 7A). Although strain differences in the ULB and TT did not reach significance, strong trends were observed. ULB strain was greater in partial and midflexion, but less in full flexion (Fig. 7B). The TT behaved opposite to the ULB (Fig. 7C).
Discussion The extensor mechanism is a complex anatomic structure that integrates the muscle forces of the finger to produce fine coordinated movements. Landsmeer observed that simultaneous motion must occur in all joints proximal to the joint under active tendon control. He attributed this phenomenon to differential loading among the components of the extensor mechanism. Subsequent to his observations several mechanical studies were performed to evaluate this intricate system, with each emphasizing particular functional characteristics. The experimental designs in these studies demonstrate the difficulties that are encountered in evaluating the extensor mechanism. For example, Sarrafian et al.' performed a strain analysis of the CB and TT, but the contribution of the intrinsics was ignored. The Hall-effect transducers used in this study made it possible to measure directly and simultaneously the strains in multiple components of the extensor mechanism with minimal tissue interference. 12 In addition, since finger positions were controlled solely by the extrinsic and intrinsic tendons, strain
7 6 5 4 3 2 1 0
1
Load 1
Load 2
~
Load 3
Figure 4. Strains in the terminal tendon as a function of tendon load in different finger positions (n=6). Error bars denote 1 SD.
The Journal of Hand Surgery / Vol. 20A No. 5 September 1995
837
7
%
s
S
4
r a
3
t
i
2
n
0 CB
/
ULB Claw
~
Hook
TT ~
Full Flexion
Figure 5. Strains in the central band (CB), ulnar lateral band (ULB), and the terminal tendon (TT) in the claw deformity. The claw deformity is compared to the hook and full flexion positions with intrinsic minus loading conditions (n=6). Error bars denote 1 SD.
7
6 5 4 3 2 1 0 -1
Central Band I
Ext
I
Terminal Tendon
Ulnar Lateral Band
CS Cut
Ext, all cut
I
Flexed, all cut
Figure 6. Strains in the central band, ulnar lateral band, and terminal tendon after each sequential step in the formation of the boutonniere deformity (see text for details) (n=6). Error bars denote 1 SD. Ext, all structures intact, finger extended; CS, central slip cut, finger extended; Ext, all cut, central slip and transverse and triangular ligaments cut, finger extended; Flexed, all cut, central slip and transverse and triangular ligaments cut, finger flexed.
838
Hurlbut and Adams / Analysis of Finger Extensor Mechanism Strains 3 2.5
% 2
S t r
1.5
a
i n
1 0.5
A
Before Deformity
~
After Deformity
Before Deformity
~
After Deformity
Before Deformity
~
After Deformity
4
% S t r
3
a
2
i n
1
B
7
%
5
S t r a i n
4
3 2
0
C
Figure 7. Strains in the (A) central band, (B) ulnar lateral band, and (C) terminal tendon as a function of finger position before and after the boutonniere deformity is created (n=6). Error bars denote 1 SD.
The Journal of Hand Surgery / Vol. 20A No. 5 September 1995 839
measurements accurately reflected the changes in tensions in each component. The effect of finger position on extensor mechanism strains was first reported by Sarrafian et al. 1 They claimed that high strains occur in the CB and T]? with full finger flexion. Although we also demonstrated higher strains in these components with flexion, the increase was much less. The different results are probably related to the difference in MPjoint position during testing. They fixed the MPjoint statically with a pin, while we allowed the joint to seek a more natural position through tendon loading. These findings emphasize the sensitivity of the extensor mechanism to finger position. In fact, the results of the strain-versus-load testing indicate that finger position is not only an influential factor, but is probably the dominant factor determining the ratio of strains within the extensor mechanism. These results also indicate that the variation in elastic properties among the extensor components does not appear to have significant clinical implications at subfailure loads. The concept that MPjoint position affects the distribution of stresses within entire extensor mechanism was noted by Bunnell and is the basis for the intrinsictightness test. 13The test compares the resistance of the PIP joint to passive flexion with the MP joint held in flexion versus when the MP joint is held in extension. The full flexion and hook positions tested in this study correspond closely to these two finger positions, respectively. ULB strain was 10 times greater in the hook position than in the full flexion position under the same loading conditions. Thus, the experimental findings strongly support the concept of the Bunnell test. There is considerable controversy concerning the anatomy and mechanics of the extensor mechanism about the PIP joint. The numerous, subtle interconnections among the extensor components in this region create confusion in anatomic descriptions. For example, there is disagreement on the proximal and distal extents of the lateral bands. We obviated part of this problem by making measurements at readily identifiable sites of the gross anatomy. However, we recognize that some of our descriptive terms may be different from those of other authors. The intricacy of the anatomy makes for a complex mechanical system that is equally difficult to delineate. Micks and Reswick proposed a mathematical model to describe the effective moment arm of the extensor mechanism around the PIP joint. 2 The model predicted that the primary extensor moment in PIP flexion is provided by the CB, while the lateral bands have the greatest extensor moment in full
PIP extension. Several investigators have observed a palmar shift of the lateral bands with PIP joint flexion. 14-16The functional significance of this shift to a position palmar to the axis of PIP joint rotation is controversial. In this location, the lateral bands would theoretically have a flexor moment. However, several authors claim that laxity develops in the lateral bands with PIP flexion, thus implying that the flexor moment would be negligible. Although the study by Sarrafian et al.' has been interpreted to imply that zero strain occurs in the lateral bands at full PIP joint flexion, these conclusions are based on relative changes in elongation and not quantitative tissue strains. In their experimental method, least elongation may not represent zero tissue strain. Our method demonstrated that strains decreased in the wing tendon region of the lateral bands during flexion but never reached zero. These findings suggest that the lateral bands have a small flexor moment across the PIP joint during finger flexion. In investigations of claw finger mechanics, Mulder 3 and Landsmeer 14''5 measured greater tension in the sagittal bands of the MP joint extensor hood and less tension in the CB. Similarly, Sarrafian et al. measured less CB strain in the claw deformity.' Our findings are in agreement, with CB strains measuring 35% less in the claw deformity than in the hook position. Reduced CB strain with MP hyperextension is attributed to an increased transfer of tension from the EDC through the sagittal bands to the palmar plate of the MPjoint. These findings support the concept of using an extension block splint or surgical procedure to prevent MP joint hyperextension in order to increase the efficiency of the EDC to extend the PIP joint. In the boutonniere deformity, disruption of the central slip causes the transfer of tension from the CB to other components of the extensor mechanism. We found that CB strain decreased after each step in creating the deformity, while ULB strain increased until the lateral bands shifted palmarly. After the palmar shift, ULB strain again decreased, but strain recovered in the TT. Since tension is theoretically transferred from the EDC to both the ulnar and radial lateral bands, the expected strain changes should be less in each lateral band than in the EDC. We found this to be true. These results are not only consistent with the observed clinical deformity, but they also emphasize the subtlety of the changes in the extensor mechanism that are responsible for the deformity. The strong correlation between the results in this study and clinical observations indicates that a strain model is an effective method of study of extensor
840
I-lurlbut and Adams / Analysis of Finger Extensor Mechanism Strains
mechanics. Because the extensor components function primarily in tension along identifiable axes, uniaxial strain gauges are well suited to the extensor mechanism. Current devices are small and require minimal exposure for attachment. The gauges do not appear to interfere with extensor function or cause gross tissue damage. In addition, since multiple gauges can be monitored simultaneously, the effects o f a finger movement, muscle force, or surgical manipulation can be measured concurrently at several sites. Thus, strains can be measured reliably and without the experimental constraints present in previous studies. Based on these observations, we believe this minimally invasive technique has potential for in vivo surgical applications in the evaluation of the extensor mechanism under dynamic conditions. We believe the results of this study strongly support Landsmeer's theory that predictable and obligatory interactions occur within the extensor mechanism during finger motion. Contemporary theories on the pathomechanics of finger deformities were also supported. In addition, the results demonstrate the delicate balance that exists in the extensor mechanism and emphasize the difficulties that surgeons encounter during their attempts to restore extensor function.
References 1. Sarrafian SK, Kazarian LE, Topouzian LK, Sarrafian VK, Siegelman A. Strain variations in the components of the extensor apparatus of the finger during flexion and extension: a biomechanical study. J Bone Joint Surg 1970;52A:980-90. 2. Micks JE, Reswick JB. Confirmation of differential loading of lateral and central fibers of the extensor tendon. J Hand Surg 1981;6:462-7. 3. Mulder JD, Landsmeer JME The mechanism of claw finger. J Bone Joint Surg 1968;50B:664-8.
4. Garcia-Elias M, An KN, Berglund LJ, Linscheid RL, Cooney WP, Chat EYS. Extensor mechanism of the fingers. II. Tensile properties of components. J Hand Surg 1991 ;16A: 1136--40. 5. Long C II. Intrinsic-extrinsic muscle control of the finger: electromyographic studies. J Biomech Eng 1978;100: 159--67. 6. Long C II, Brown ME. Electromyographic kinesiology of the hand: muscles moving the long finger. J Bone Joint Surg 1964;46A: 1683-705. 7. Long C It, Conrad PW, Hall EA, Fufler SL. Intrinsicextrinsic muscle control of the hand in power grip and precision handling: an electromyographic study. J Bone Joint Surg 1970;52A:853-67. 8. Long C II, Hall EA. Intrinsic hand muscles in power grip. Electromyography 1968;8:397--421. 9. Zancolli EA. Structural dynamic bases of hand surgery. 2nd ed. Philadelphia: Raven Press, 1979:79-91,159-67. 10. Mason RR, Munro RR. Relationship between amplitude of EMG potentials and tension in abduction of the little finger. J Anat 1970;106:185-99. 11. Brand PW, Beach RB, Thompson DE: Relative tension and potential excursion of muscles in the forearm and hand. J Hand Surg 198l;6:209-19. 12. Arms SA, Pope MH, Johnson RJ, Fischer RA, Arvidsson I, Eriksson E. The biomechanics of anterior cruciate ligament rehabilitation and reconstruction. Am J Sports Med 1984;12:8-18. 13. Boyes JH. Bunnel's surgery of the hand. 4th ed. Philadelphia: JB Lippincott, 1964:256-60. 14. Landsmeer JMF, Long C. The mechanism of finger control, based on electromyograms and location analysis. Acta Anat 1965;60:330-47. 15. Landsmeer JMF. The anatomy of the dorsal aponeurosis of the human finger and its functional significance. Anat Rev 1949;104:31-43. 16. Garcia-Elias M, An KN, Berglund L, Linscheid RL, Cooney WP, Chat EYS. Extensor mechanism of the fingers. I. A quantitative geometric study. J Hand Surg 1991;16A:1130--36.
The extensor mechanism of the finger is an elaborate system that integrates the forces of the lumbrical, interossei, and long extensor to produce precise functional movements. Although detailed descriptions of the functional anatomy have been reported, biomechanical support for current theories on extensor mechanics is limited. Previous studies have focused primarily on specific intricacies of the extensor mechanism. A strain analysis of the extensor mechanism was performed by Sarrafian et al.; however, the contribution of the intrinsics was not included. 1 Micks and Reswick developed a mathematical model based on the effective moment arms of the extensor components about the proximal interphalangeal (PIP) joint. 2 The model predicted strains in the central slip and lateral band but considered only PIP joint motion. Mulder and Landsmeer investigated extensor mechanism strains about the metacarpophalangeal (MP) From the University of Vermont, Depaltment of Orthopaedics and Rehabilitation, Burlington, VT and the University of Iowa, Department Orthopaedic Surgery, Iowa City, IA. Received for publication March 21, 1994; accepted in revised form March 17, 1995. 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. Reprint requests: Brian D. Adams, MD, Associate Professor, Division of Hand and Microsurgery, Department Orthopaedic Surgery, University of Iowa Hospitals & Clinics, Iowa City, IA 52242.
832
The Journal of Hand Surgery
joint associated specifically with the claw deformity. 3 In a study of geometric alignment and tensile properties of the extensor mechanism, Garcia-Elias et al. demonstrated significant variation among components, but deformities were not considered? Defining the interaction of forces within the extensor mechanism is important to the understanding of normal extensor function as well as the pathomechanics of many finger deformities. The goal of this study was to test contemporary theories on extensor mechanics of the normal and pathologic finger using a tissue strain model. Strain is a measure of tissue deformation resulting from an applied force. Although strain is an indirect indicator of force, it can be measured, unlike load, with minimal disturbance of tissues. To avoid the limitations of previous studies, an experiment was designed to evaluate the extensor mechanism in different composite finger positions and with different intrinsic and extrinsic tendon load conditions. During this experiment, we sought to assess the feasibility of using this method for in vivo evaluation of the extensor mechanism.
Materials and Methods Twelve fresh-frozen human cadaver long fingers amputated at the level of the proximal metacarpal were used. Specimens were thawed prior to testing and kept moist with normal saline. Dorsal skin and subcuta-
The Journal of Hand Surgery / Vol. 20A No. 5 September 1995
neous tissues were removed. A 1/8-inch steel rod was inserted into the medullary canal of the metacarpal and fixed with bone cement. The construct was mounted in a testing jig that allowed the MP and interphalangeal (IP) joints free range of motion. Fingers were tested in six different positions with composite angles of IP and MP joint flexion set at extension (0~ partial flexion (90~ midflexion (180~ full flexion (260~ intrinsicplus (60~ and hook (180 ~ (Fig. 1). To achieve these positions, monofilament nylon lines were attached to the tendons of the extensor digitorum communis (EDC), interossei, and lumbricals and routed over low friction pulleys to free weights. The proximal ends of the flexor digitorum superficialis and profundus tendons were clamped at their resting lengths in each of the six positions. Although an infinite number of tendon load combinations could potentially achieve a given finger position, previously reported electromyographic (EMG) data were used to determine muscle activity for each specific finger position?-8 Using this information, the following constraints were applied: (1) lumbrical activity is always present with IP joint extension,
regardless of MP joint position and (2) interossei have their greatest activity during IP joint extension with simultaneous MPjoint flexion. In addition, some interossei force is present in the fully flexed and extended positions as well as during finger flexion to provide smooth synchronous motion of all joints. However, intrinsic activity is absent in the claw d e f o r m i t y . 9 Since EMG data correlate poorly with muscle cell recruitment and do not predict tension magnitude,~~ the tension fractions of the muscles were used to specify the magnitudes of applied loads, u Three loading conditions of increasing weight were used. In the first loading condition, the loads used for the active muscles in a given finger position were set at 5% of their maximum predicted tensions (MPT), with some variations to satisfy the other experimental conditions (Fig. 1). The second and third loading conditions were determined by multiplying the percentage of MPT used in the first loading condition by 2 and 5 times, respectively. Thus, loads in the second condition were approximately 10% of MPT and those in the third condition were approximately 25% of MPT. The two
F~C
Finger P ~ i t t o n
c:::=::t~:::x::~
833
tl
Composite AngLe
Loed Conditi~'~
~PT
LOOd(e)
Ext4mlton
I
$
475
2
10
3
25
~g4PY
UI
L
Loadtg)
NEPT
Loed(g)
INPT
Z
250
2
~O
15
Loci(g) 150,
950
4
500
~
1,00
30
300
:X73
10
1250
10
1000
75
nO
O"
Portia( FLexion
:
r~ I
I
;-~,I
:
:
I~
I
C,I
i
i n u n u u l i I
90"
I;I~
IJ,
t*,I
I I , I
I
I'~-~I ,
I I , I
I[
I
I
,:,:,I
5
675
2
250
2
200
2
10
950
&
500
6
400
5 5
3
25
2375
10
1250
10
1000
5
Ftexion 180"
FuLL
1
5
~?~
2
25o
2
~
5
Ftexion 260"
2
10
950
4
500
4
40Q
5
3
25
z375
10
1~$0
10
100Q
5
Intrtnelr Ptul
Nook 180"
1
2.5
237.5
5
625
5
500
15
150
2
5
475
10
~250
10
3000
30
300
3
12.5
1~87.5
25
3125
25
25~
75
730
1
5
4?3
8
9
5
2
10
950
8
9
5
$
25
1187.5
8
9
5
Figure 1. Loading conditions for each corresponding finger position. %MPT, percentage of maximum predicted tension; EDC, extensor digitorum communis; RI, radial interosseous; UI, ulnar interosseous; L, lumbrical; Load(g), load in grams.
834 HurlbutandAdams/ Analysisof FingerExtensorMechanismStrains higher loads were applied for two purposes: first, to assure a response of the strain gauges to changes in load, and second, to determine if the ratio of strains among the extensor components varies with applied loads due to differences in the elastic properties among the components. All experimental conditions were satisfied when (1) EMG constraints were met, (2) the first loading condition held the finger in the desired position, and (3) finger position was maintained when weight was increased to the second and third loading conditions (Fig. 1). Strains measured under these conditions will reflect total force resulting from applied load and passive stretch. For example, in the intrinsicplus position there are high intrinsic forces but least passive stretch acting on the lateral bands. The opposite occurs in the hook position. Extensor mechanism strains were measured using Hall-effect transducers that were customized for this application (Micro Strain Inc., Burlington, VT). The gauges were attached to the extensor mechanism by short barbs on the ends of the device and 5-0 nylon suture (Fig. 2). Attachment was simple, quick, and required only 4 to 8 mm of tissue for each gauge. The
gauges did not cause tissue buckling, twisting, or other signs of interference with extensor function. The barbs and sutures held the gauges secure during testing and there was no grossly identifiable tissue damage after their removal. Gauges were mounted on the extensor mechanism after nominal loads (5 grams) were applied to the intrinsic and extrinsic tendons to remove observable slack from the tendons and lines. Baseline lengths measured at nominal loads were used for strain calculations. Regression analysis was performed to determine instances of significant specimen effect. Comparisons were derived using Fisher's least significant difference test and Bonferroni's multiple comparisons procedure to establish significant differences among conditions. Bonferroni's procedure was included to accommodate for the high number of comparisons. Least squares means were calculated to adjust for occasional missing data points. SAS software was used for statistical analysis with statistical significance taken at the 95% level (p < .05). Each specimen was tested using a combination of two of the following three protocols.
~
HII ~ Tmd~sduow Termlnll Two:Ion
d Exmnsor D - - m
//
Comrnunlm
//
Hall ~ TmnKl~[mr FIBdlIII ~ 1
\ Figure 2. Radial view of finger showing the attachment sites for three of the Hall-effect transducers. The ulnar lateral band implantation site (not illustrated) corresponds to a location similar to that shown for the radial lateral band. The fingers were mounted in a specially designed jig (bottom left).
The Journal of Hand Surgery / Vol. 20A No. 5 September 1995 835 Strain Distribution versus Finger Position In all 12 specimens, strains were measured at three sites on the intact extensor mechanism: the central band (CB) of the EDC tendon proximal to its junction with the oblique fibers from the lateral bands, the ulnar lateral band (ULB) proximal to its junction with the oblique fibers from the central band, and the terminal tendon (TT) distal to the triangular ligament (Fig. 2). Testing was done in all six finger positions using the second loading condition.
Second, the triangular and transverse retinacular ligaments were cut parallel with the fibers of the lateral bands. In the final step, the PIP joint was flexed to cause palmar subluxation of the lateral bands. Strains were measured after each step with the finger in full extension. After testing in full extension, the digit was tested in partial flexion, midflexion, full flexion, and the intrinsic-minus position.
Results Strain Distribution versus Finger Position
Strain Distribution versus Tendon Load In 6 of 12 specimens, strain measurements were made for all three loading conditions. In addition to the measurements made on the CB, ULB, and TT, strains were measured on the radial lateral band (RLB) proximal to its junction with the oblique fibers from the central slip (Fig. 2). Testing was done in all six finger positions. Claw and Boutonni?~re Deformities In the remaining six specimens, the full flexion position, hook position, and claw deformity were tested with no loads applied to the interossei and lumbrical tendons and 10% of MPT applied to the EDC tendon. Strains were measured at three locations (CB, ULB, and TT). A boutonniere deformity was then created using the sequence of cuts described by Zancolli. 9First, the central slip of the EDC tendon was cut near its attachment to the proximal phalanx.
Mean strain in the CB increased with increasing finger flexion, with 64% greater strain at full flexion. CB strain in midflexion, full flexion, and the hook position was significantly greater than in full extension (p < .05). Greatest CB strain occurred in the hook position and at full flexion. ULB strain decreased with increasing finger flexion, with 40% less strain in full flexion. Greatest ULB strain occurred at extension and in the hook position. Least ULB strain occurred in the intrinsic-plus position. ULB strain in full flexion and in the intrinsic-plus position was significantly less than in full extension (p < .05). TT strain increased with increasing finger flexion, with 124% greater strain in full flexion than in full extension. Greatest TT strain occurred in full flexion and in the hook position. Partial flexion, midflexion, full flexion, and the hook position caused significantly greater TT strain than occurred in extension and the intrinsic-plus position (p < .05) (Fig. 3).
6 5
% 4
r a n
Central Band
m~] Ulnar Lateral Band
[
I Terminal Tendon
Figure 3. Strains in the central band, terminal tendon, and ulnar lateral band as a function of finger position (n= 12). Error bars denote 1 SD.
836
Hurlbut and Adams / Analysis of Finger ExtensorMechanism Strains
Strain Distribution versus Tendon Loads
Increasing loads on the tendons did not alter the relative strain distribution for each finger position. Strains increased progressively in the CB, RLB, ULB, and TT with heavier load conditions. For example, Figure 4 demonstrates the changes that occur in the TT with increasing load. Claw and Boutonniere Deformities
Mean strains in the claw deformity, hook position, and full flexion position were evaluated for significant differences. CB strain was less in the claw deformity than in full flexion (p < .05) or the hook position (p = .07). ULB strain was greater in the hook position than in the claw deformity or in full flexion (p < .05); in fact, ULB strain was 10 times greater in the hook position than in full flexion (p < .05). No significant differences in TT strain occurred among the finger positions (Fig. 5). CB strain decreased with each step during creation of the boutonniere deformity (Fig. 6). ULB strain increased slightly after the central slip was cut and again increased slightly after the triangular and transverse retinacular ligaments were cut. ULB strain then decreased slightly after the lateral bands migrated palmarly. TT strain decreased slightly with the first cut but recovered after the lateral bands migrated palmarly. Strains in each complete boutonniere deformity were compared to those in the intact digit. CB strain in the boutonnibre deformity averaged 60%
less for all finger positions (p < .05) (Fig. 7A). Although strain differences in the ULB and TT did not reach significance, strong trends were observed. ULB strain was greater in partial and midflexion, but less in full flexion (Fig. 7B). The TT behaved opposite to the ULB (Fig. 7C).
Discussion The extensor mechanism is a complex anatomic structure that integrates the muscle forces of the finger to produce fine coordinated movements. Landsmeer observed that simultaneous motion must occur in all joints proximal to the joint under active tendon control. He attributed this phenomenon to differential loading among the components of the extensor mechanism. Subsequent to his observations several mechanical studies were performed to evaluate this intricate system, with each emphasizing particular functional characteristics. The experimental designs in these studies demonstrate the difficulties that are encountered in evaluating the extensor mechanism. For example, Sarrafian et al.' performed a strain analysis of the CB and TT, but the contribution of the intrinsics was ignored. The Hall-effect transducers used in this study made it possible to measure directly and simultaneously the strains in multiple components of the extensor mechanism with minimal tissue interference. 12 In addition, since finger positions were controlled solely by the extrinsic and intrinsic tendons, strain
7 6 5 4 3 2 1 0
1
Load 1
Load 2
~
Load 3
Figure 4. Strains in the terminal tendon as a function of tendon load in different finger positions (n=6). Error bars denote 1 SD.
The Journal of Hand Surgery / Vol. 20A No. 5 September 1995
837
7
%
s
S
4
r a
3
t
i
2
n
0 CB
/
ULB Claw
~
Hook
TT ~
Full Flexion
Figure 5. Strains in the central band (CB), ulnar lateral band (ULB), and the terminal tendon (TT) in the claw deformity. The claw deformity is compared to the hook and full flexion positions with intrinsic minus loading conditions (n=6). Error bars denote 1 SD.
7
6 5 4 3 2 1 0 -1
Central Band I
Ext
I
Terminal Tendon
Ulnar Lateral Band
CS Cut
Ext, all cut
I
Flexed, all cut
Figure 6. Strains in the central band, ulnar lateral band, and terminal tendon after each sequential step in the formation of the boutonniere deformity (see text for details) (n=6). Error bars denote 1 SD. Ext, all structures intact, finger extended; CS, central slip cut, finger extended; Ext, all cut, central slip and transverse and triangular ligaments cut, finger extended; Flexed, all cut, central slip and transverse and triangular ligaments cut, finger flexed.
838
Hurlbut and Adams / Analysis of Finger Extensor Mechanism Strains 3 2.5
% 2
S t r
1.5
a
i n
1 0.5
A
Before Deformity
~
After Deformity
Before Deformity
~
After Deformity
Before Deformity
~
After Deformity
4
% S t r
3
a
2
i n
1
B
7
%
5
S t r a i n
4
3 2
0
C
Figure 7. Strains in the (A) central band, (B) ulnar lateral band, and (C) terminal tendon as a function of finger position before and after the boutonniere deformity is created (n=6). Error bars denote 1 SD.
The Journal of Hand Surgery / Vol. 20A No. 5 September 1995 839
measurements accurately reflected the changes in tensions in each component. The effect of finger position on extensor mechanism strains was first reported by Sarrafian et al. 1 They claimed that high strains occur in the CB and T]? with full finger flexion. Although we also demonstrated higher strains in these components with flexion, the increase was much less. The different results are probably related to the difference in MPjoint position during testing. They fixed the MPjoint statically with a pin, while we allowed the joint to seek a more natural position through tendon loading. These findings emphasize the sensitivity of the extensor mechanism to finger position. In fact, the results of the strain-versus-load testing indicate that finger position is not only an influential factor, but is probably the dominant factor determining the ratio of strains within the extensor mechanism. These results also indicate that the variation in elastic properties among the extensor components does not appear to have significant clinical implications at subfailure loads. The concept that MPjoint position affects the distribution of stresses within entire extensor mechanism was noted by Bunnell and is the basis for the intrinsictightness test. 13The test compares the resistance of the PIP joint to passive flexion with the MP joint held in flexion versus when the MP joint is held in extension. The full flexion and hook positions tested in this study correspond closely to these two finger positions, respectively. ULB strain was 10 times greater in the hook position than in the full flexion position under the same loading conditions. Thus, the experimental findings strongly support the concept of the Bunnell test. There is considerable controversy concerning the anatomy and mechanics of the extensor mechanism about the PIP joint. The numerous, subtle interconnections among the extensor components in this region create confusion in anatomic descriptions. For example, there is disagreement on the proximal and distal extents of the lateral bands. We obviated part of this problem by making measurements at readily identifiable sites of the gross anatomy. However, we recognize that some of our descriptive terms may be different from those of other authors. The intricacy of the anatomy makes for a complex mechanical system that is equally difficult to delineate. Micks and Reswick proposed a mathematical model to describe the effective moment arm of the extensor mechanism around the PIP joint. 2 The model predicted that the primary extensor moment in PIP flexion is provided by the CB, while the lateral bands have the greatest extensor moment in full
PIP extension. Several investigators have observed a palmar shift of the lateral bands with PIP joint flexion. 14-16The functional significance of this shift to a position palmar to the axis of PIP joint rotation is controversial. In this location, the lateral bands would theoretically have a flexor moment. However, several authors claim that laxity develops in the lateral bands with PIP flexion, thus implying that the flexor moment would be negligible. Although the study by Sarrafian et al.' has been interpreted to imply that zero strain occurs in the lateral bands at full PIP joint flexion, these conclusions are based on relative changes in elongation and not quantitative tissue strains. In their experimental method, least elongation may not represent zero tissue strain. Our method demonstrated that strains decreased in the wing tendon region of the lateral bands during flexion but never reached zero. These findings suggest that the lateral bands have a small flexor moment across the PIP joint during finger flexion. In investigations of claw finger mechanics, Mulder 3 and Landsmeer 14''5 measured greater tension in the sagittal bands of the MP joint extensor hood and less tension in the CB. Similarly, Sarrafian et al. measured less CB strain in the claw deformity.' Our findings are in agreement, with CB strains measuring 35% less in the claw deformity than in the hook position. Reduced CB strain with MP hyperextension is attributed to an increased transfer of tension from the EDC through the sagittal bands to the palmar plate of the MPjoint. These findings support the concept of using an extension block splint or surgical procedure to prevent MP joint hyperextension in order to increase the efficiency of the EDC to extend the PIP joint. In the boutonniere deformity, disruption of the central slip causes the transfer of tension from the CB to other components of the extensor mechanism. We found that CB strain decreased after each step in creating the deformity, while ULB strain increased until the lateral bands shifted palmarly. After the palmar shift, ULB strain again decreased, but strain recovered in the TT. Since tension is theoretically transferred from the EDC to both the ulnar and radial lateral bands, the expected strain changes should be less in each lateral band than in the EDC. We found this to be true. These results are not only consistent with the observed clinical deformity, but they also emphasize the subtlety of the changes in the extensor mechanism that are responsible for the deformity. The strong correlation between the results in this study and clinical observations indicates that a strain model is an effective method of study of extensor
840
I-lurlbut and Adams / Analysis of Finger Extensor Mechanism Strains
mechanics. Because the extensor components function primarily in tension along identifiable axes, uniaxial strain gauges are well suited to the extensor mechanism. Current devices are small and require minimal exposure for attachment. The gauges do not appear to interfere with extensor function or cause gross tissue damage. In addition, since multiple gauges can be monitored simultaneously, the effects o f a finger movement, muscle force, or surgical manipulation can be measured concurrently at several sites. Thus, strains can be measured reliably and without the experimental constraints present in previous studies. Based on these observations, we believe this minimally invasive technique has potential for in vivo surgical applications in the evaluation of the extensor mechanism under dynamic conditions. We believe the results of this study strongly support Landsmeer's theory that predictable and obligatory interactions occur within the extensor mechanism during finger motion. Contemporary theories on the pathomechanics of finger deformities were also supported. In addition, the results demonstrate the delicate balance that exists in the extensor mechanism and emphasize the difficulties that surgeons encounter during their attempts to restore extensor function.
References 1. Sarrafian SK, Kazarian LE, Topouzian LK, Sarrafian VK, Siegelman A. Strain variations in the components of the extensor apparatus of the finger during flexion and extension: a biomechanical study. J Bone Joint Surg 1970;52A:980-90. 2. Micks JE, Reswick JB. Confirmation of differential loading of lateral and central fibers of the extensor tendon. J Hand Surg 1981;6:462-7. 3. Mulder JD, Landsmeer JME The mechanism of claw finger. J Bone Joint Surg 1968;50B:664-8.
4. Garcia-Elias M, An KN, Berglund LJ, Linscheid RL, Cooney WP, Chat EYS. Extensor mechanism of the fingers. II. Tensile properties of components. J Hand Surg 1991 ;16A: 1136--40. 5. Long C II. Intrinsic-extrinsic muscle control of the finger: electromyographic studies. J Biomech Eng 1978;100: 159--67. 6. Long C II, Brown ME. Electromyographic kinesiology of the hand: muscles moving the long finger. J Bone Joint Surg 1964;46A: 1683-705. 7. Long C It, Conrad PW, Hall EA, Fufler SL. Intrinsicextrinsic muscle control of the hand in power grip and precision handling: an electromyographic study. J Bone Joint Surg 1970;52A:853-67. 8. Long C II, Hall EA. Intrinsic hand muscles in power grip. Electromyography 1968;8:397--421. 9. Zancolli EA. Structural dynamic bases of hand surgery. 2nd ed. Philadelphia: Raven Press, 1979:79-91,159-67. 10. Mason RR, Munro RR. Relationship between amplitude of EMG potentials and tension in abduction of the little finger. J Anat 1970;106:185-99. 11. Brand PW, Beach RB, Thompson DE: Relative tension and potential excursion of muscles in the forearm and hand. J Hand Surg 198l;6:209-19. 12. Arms SA, Pope MH, Johnson RJ, Fischer RA, Arvidsson I, Eriksson E. The biomechanics of anterior cruciate ligament rehabilitation and reconstruction. Am J Sports Med 1984;12:8-18. 13. Boyes JH. Bunnel's surgery of the hand. 4th ed. Philadelphia: JB Lippincott, 1964:256-60. 14. Landsmeer JMF, Long C. The mechanism of finger control, based on electromyograms and location analysis. Acta Anat 1965;60:330-47. 15. Landsmeer JMF. The anatomy of the dorsal aponeurosis of the human finger and its functional significance. Anat Rev 1949;104:31-43. 16. Garcia-Elias M, An KN, Berglund L, Linscheid RL, Cooney WP, Chat EYS. Extensor mechanism of the fingers. I. A quantitative geometric study. J Hand Surg 1991;16A:1130--36.