Journul (fOrthupuedrr Rewarch 10:157-166 Raven Press, Ltd , New York

C 1992 Orthopaedic Research Society

The Early Effects of Joint Immobilization on Medial Collateral Ligament Healing in an ACL-Deficient Knee: A Gross Anatomic and Biomechanical Investigation in the Adult Rabbit Model "R. C. Bray, "TN. G. Shrive, "C. B. Frank, and t D . D. Chimich Joint lnjury und Diseases Research Group, Faculty of Medicine, *Department of Surgery and +Department of Civil Engineering, University uf Calgary, Calgory, Alberta, Canada

Summary: In this study, the short-term effects of immobilization on joint damage and medial collateral ligament (MCL) healing were investigated in unstable, anterior cruciate ligament (ACL)-deficient knees in rabbits. Forty-six 12month-old female New Zealand white rabbits were separated into three groups. Animals from each group had surgery on their right knees: group I, sham controls ( n = 9 ) ; group 11, complete transection of the ACL and removal of a 4 mm segment (gap injury) of MCL midsubstance with no immobilization of the limb (n = 19): and group 111, same injuries to the ACL and MCL (as group 11) but with immobilization of the limb (n = 18). No surgical repair of disrupted ligaments was performed. Left knees served as unoperated contralateral controls. All animals were allowed unrestricted cage activity until sacrifice in subgroups at 3 , 6 , and 14 weeks of healing when biomechanical properties of all MCLs were measured. All knee joints were systematically examined for gross evidence of damage to articular cartilage, menisci, and periarticular soft tissues. To monitor relative in vivo loads on injured limbs during healing, hindlimb weight bearing was assessed at biweekly intervals. Results indicated that animals in both groups I1 and I11 bore relatively lower loads (compared to preinjury values) on their injured hindlimbs. Mechanical testing of MCLs showcd only minor changes in sham controls, while group I1 and I11 healing MCLs demonstrated significantly lower force and stress at MCL complex failure compared to contralateral controls. In specific comparisons of group I11 to group I1 animals, we noted that immobilization prevented joint damage over the early intervals studied. In addition, immobilization resulted in MCL laxity similar to contralateral control values but inhibited development of structural strength and stiffness in healing MCLs. These results suggest that in the rabbit, short-term immobilization of an ACL-deficient knee offers some advantages to the joint and to certain low load behaviors of the healing MCL, but it also results in a smaller quantity of scar tissue that is less able to resist higher loads. Longcr-term studies involving remobilization are necessary before the effects of brief immobilization on joint damage and MCL healing in this ACL-deficient model can be fully defined. Key Words: Medial collateral ligament-HealingImmobilization-Anterior cruciate ligament-Rabbits.

Joint injury associated with multiple ligament disruptions can potentially lead to long-term instability and ProgressiveJoint damage (11). In the knee, such severe injuries are often treated by ligament repair

Received January 2, 1991; accepted September 19, 1991. Address correspondence and reprint requests to Dr. R. c. Bray at Department of Surgery, The University of Calgary, Health Sciences Centre, 3330 Hospital Drive N . w . ,Calgary, Alberta, Canada, T2N 4N1.

157

158

R. C . BRAY ET AL.

or joint immobilization (9) in the hope that ligament healing will be improved and joint damage will be minimized (6,14,15). Hence, the empirical rationale for immobilization seems to be to protect healing tissues from abnormal stresses and to prevent further damage to healing and normal tissues in a mechanically unstable joint (6,22). Many clinical studies of ligament healing have therefore recommended at least some form of short-term joint immobilization as treatment (9,15,19). While the intended benefits of immobilization are appealing, the desirable effects are not always obtained. For example, in normal joints, immobilization simultaneously induces disuse atrophy of healthy structures such as tendons, ligaments, and bone (1). In animal models, normal articular cartilage shows evidence of damage after relatively short (6 week) periods of immobilization (16). The healing of certain ligaments is also thought to be adversely affected by immobilization (12,13). Thus, the benefits of immobilizing an injured joint must be balanced against these potentially detrimental effects. Severe injuries with multiple ligament disruptions and concomitant joint instability, however, are thought to benefit from short-term immobilization based on evidence that joint function and ligament healing (15) are improved (when remobilization is instituted relatively quickly). Woo and co-workers (22) as well as Hart and Dahners (8) studied injuries to the medial collateral ligament (MCL) of the knee in different animal models of severe joint injury. There was significant instability if the anterior CI-Uciate ligament (ACL) was disrupted in both of these models. Nonimmobilized knees had greater than normal valgus rotation and had joint damage. One conclusion from these studies was that treatment was required to prevent gross displacements of the knee and presumably to protect the healing MCL. We hypothesized that immobilizing an unstable knee should prevent gross joint damage and simultaneously permit adequate ligament healing to occur. To test this hypothesis, the investigation described here was carried out in a previously defined rabbit model of MCL healing (3.7). Joint instability was induced by transection of the ACL (2). A controlled gap injury was created in the MCL (4)and MCL healing in ACL-deficient knees with or without immobilization was assessed at three short-term intervals. Gross evidence of joint damage was recorded in all specimens. Mechanical tests were per-

J Orthop Res, Vol. 10,No. 2 , 1992

formed to quantify how MCL healing in an ACLdeficient knee was affected by immobilization. MATERIALS AND METHODS

Forty-six adult, female, New Zealand white rabbits (aged 12 months) weighing between 3.6 and 6.4 kg (mean of 4.7 2 0.66 kg) were used (Riemens Fur Ranches Ltd., St. Agatha, Ontario, Canada). All animals had surgery under sterile conditions using a general inhalational anaesthetic (1% halothane and oxygen at 1 Limin). Only right knees were operated on leaving the left knees as contralateral controls. These “control” knees were not considered normal due to possible effects of the opposite knee surgery. After surgery, animals were allowed unrestricted activity in cages until sacrifice at 3 , 6 , and 14 weeks. All animals received similar diets postoperatively [Teklad Rabbit Ration (Doghouse Kennel Supply Ltd., Edmonton, Alberta, Canada) and water ad libitum]. Study Groups In group I, nine animals underwent a sham operation (three for each sacrifice interval). Longitudinal skin and fascia1 incisions were made over the medial side of the right knee. The MCL was completely exposed by incising the anterior and posterior borders of the MCL with a #15 scalpel. The attachment of the MCL to the medial meniscus was severed but the MCL itself was not disrupted. The medial incision also permitted exposure of the lateral joint capsule by retracting the skin incision laterally, where an anterolateral capsulotomy between the tendon of extensor digitorum longus and the patellar tendon allowed direct exposure of the ACL. The ACL was then simply probed with a blunt hook. The joint capsule was closed with 6-0 nylon suture. The medial investing fascia was left to cover the exposed MCL. Skin incisions were closed with interrupted 4-0 nylon sutures and a dressing of gentian violet was applied to the wound. In group 11, 19 animals (sets of 6, 7, and 6, for the three sacrifice intervals) had a similar operation on their right knees to expose the MCL and ACL. The MCL in this group was injured by creating a 4 mm gap in the ligament midsubstance centered at the medial joint line (4) (Fig. 1). Four 6-0 nylon marking sutures were placed at the four corners of the MCL gap injury site to assist later in identification of scar

JOINT IMMOBILIZATION AND MCL HEALING

159

described (18). Ligaments were then cut as described in group IT. Animals in group I11 were defined as “immobilized” if they showed no voluntary joint motion during cage activity and less than a few degrees of passive motion on manual examination.

In Vivo Tests of Hindlimb Load Distribution

1 Ill I

v

MCL/ACL Combined Non-Immobilized GROUP II

MCL/ACL Combined Pin-Immobilized GROUP lU

FIG. 1. Schematic diagram of knee joints depicting experimental MCL and ACL disruptions performed in group I1(nonimmobilized) and group 111 (pin immobilized) animals. Lateral collateral and posterior cruciate ligaments are not illustrated.

tissue. A 4 mm gap segment of MCL was removed to insure that no contact of cut ligament ends would occur in healing MCLs and for comparison with previous studies of MCL gap healing (4). The ACL, approached through the lateral capsulotomy incision, was hooked with a blunt probe and transected completely with a #12 curved scalpel. All ACL transections were confirmed by subsequent probing of the intercondylar notch. Any uncut ACL fibers were identified and disrupted. No repair of the ACL or MCL was performed and wounds were closed as in group I. Pilot studies of combined ligament disruptions in this model had revealed that malalignments of the femur and tibia frequently occurred if ligament disruptions preceded immobilization. Since this would clearly limit the reproducibility of knee and MCL alignment, knees were immobilized prior to cutting any ligaments in group 111 animals. The 18 group 111 animals, therefore, had their right knee immobilized at approximately 150-160” of flexion by means of an extra-articular 1.6 mm stainless steel transfixing pin (Fig. 1) placed as previously

In order to determine the effect of experimental knee injury on “in vivo” static hindlimb load bearing, preoperative and weekly postoperative sitting hindlimb load distributions were obtained. A force platform consisting of multiple pressure sensors (EMED Corporation, Model F01, Munich, West Germany), interfaced with a computer (Hewlett Packard Model #7475A, Hewlett Packard Co, Palo Alto, CA, U.S.A.), was used to measure relative load distribution between the right and left hindlimbs. In each test, the animal was placed on a 250 x 350 mm force platform and allowed to assume a normal sitting position. The mean paw pressure and total contact area were measured to assess load on the right and left hind paws. Measurements consisted of five sequential trials of hindlimb load bearing when animals were sitting quietly. The five results were averaged to compute hindlimb load distribution, which was expressed as a ratio of right hindlimb load over total hindlimb load (right + left). The right hindlimb was chosen as the index hindlimb since it would eventually become the injured extremity. Thirty-two animals were selected randomly and tested on the force platform prior to surgery. The results were used to establish a normal baseline for subsequent comparison to postoperative values. A minimum of five animals were tested per interval from groups I1 and 111, but only two from group I (sham). Gross Assessments of Joint Damage

Animals were sacrificed with an overdose of sodium pentobarbital (Euthanyl, 2.0 mU4.5 kg, M.T.C. Pharmaceutical, Canada Packers Inc., Cambridge, Ontario, Canada). Hindlimbs were disarticulated immediately at the hip, leaving adjacent skin and soft tissues intact. The left limb was placed in an airtight bag and tested within 4 h of animal sacrifice. In preparation for mechanical testing, all

J Orthop Res, Vol. 10, No. 2 , 1992

R . C.BRAY ET AL.

160

knee joints were dissected carefully and assessed systematically for evidence of gross tissue damage. The presence of fluid within the joint cavity, erosions of articular cartilage, proliferation of tissue at joint margins, damage to menisci, and thickening of periarticular soft tissues (synovium and joint capsule) were recorded for each test specimen. Biomechanical Testing The combined MCLiACL disruption resulted in dramatic and persistent joint instability with considerable potential for variation in the position of the tibia relative to the femur when specimens were mounted in the test clamps. A reproducible tibiofemoral test position was therefore established by means of a specially designed joint alignment device (Fig. 2) that was attached directly to the tibial clamp. The alignment device consisted of a rectangular plexiglass plate fixed to the tibial clamp so that the plate abutted the tibial tubercle and was oriented tangentially to the femoral condyles. The plexiglass plate was fixed to the femur with an elastic band, causing contact alignment of the two femoral condyles and the tibial tubercle at three points on the surface of this plate. In a pilot study, these contact points had been determined to occur at about 70” of flexion in a group of normal knees (using the same tibial clamp in each test). The femur was then cemented with polymethylmethacrylate (PMMA) in a second clamp that freely accommodated the transected femoral bone end, and was at-

tached to the testing machine base below. The knee joints were ready for testing when the PMMA had cured, and the alignment device had been removed. Both menisci, the posterior cruciate ligament (PCL), and the lateral collateral ligament (LCL) were then removed, isolating the bone-MCGbone complex for mechanical testing. Each MCL complex was subjected to a testing sequence (4) on an Instron 1122 materials testing machine (Tnstron Corp., Canton, MA, U.S.A.). MCL “laxity” and midsubstance cross sectional area (CSA) (4) were measured first. In this and previous studies from our laboratory, MCL laxity is defined as the distance (of crosshead movement) between joint compression and the position at which the MCL first resists tensile distraction during a low-load compression-tension cycle (4). In this definition of MCL laxity, low load limits ofjoint compression and MCL tension were arbitrarily chosen to provide a useful measure of the low load extensibility of the MCL and to minimize any potential load-induced damage to the MCL test complex. A cyclic and static tensile conditioning loading series (to a fixed displacement of 0.68 mm at a rate of 10 mmimin) was then performed followed by tensile distraction to failure at an extension rate of 20 mmlmin. All data were collected on line at 50 Hz with an HP6944 Multiprogrammer (Hewlett Packard Co.). Load, deformation, and ligament midsubstance surface strain (20) were recorded simultaneously during the failure test, and the mode of failure (midsubstance, insertional, and bone avulsion) was also noted.

FIG. 2. Schematic of frontal and sagittal views of a test complex mounted in clamps (D and E) of a materials-testing machine. An alignment device is used to improve the reproducibility of joint and MCL alignment prior to testing. Before cementing the femur (F), a plexiglass plate (A) was placed to align three coplanar points [l on the tibial tubercle, and 2, 3 on each femoral condyle (insets)]. Contact alignment of the two bones through these three points resulted in a similar joint configuration for each test. The alignment device is adjustable with horizontal (B) and vertical (C) axes of motion. The alignment device is removed after cementing bones in the test clamps.

J Orthop Rrs, Voi. 10, No. 2 , 1992

JOINT IMMOBILIZATION A N D MCL HEALING Statistical Analysis All analyses were carried out on a Honeywell VIP7400 computer (Honeywell Inc., Minneapolis, MN, U.S.A.) using the SPSS and BMDP (PZV) statistical packages (5). Data analysis consisted of analysis of variance (ANOVA) followed by the Student-Neumdnn-Keul’s procedure for multiple comparisons. Significance was accepted at an alpha level of less than 0.05. Analyses were performed on data from sham, experimental, and contralateral MCL complcxes for each time interval ( 3 , 6, and 14 weeks). Data from all contralateral MCL complexes were pooled independently at each healing interval. Each experimental group (I, 11, and 111) was analyzed against the others as well as with pooled contralaterals at each interval. Differences between experimental (right) and pooled contralateral (left) MCL complexes were used to assess effects of surgical injury, joint immobilization, and time after operation on MCL healing. No statistical analyses were performed on the subjective assessments of joint damage.

I

161

GROUP1

T

0.0

I

GRP I1 AT ANY INTERVAL I

PRE-OP

GROUP11

-0- GROUP111

pr0.05 I

0 POST-OP 5

I

10

15

WEEKS

FIG. 3. Plot of in vivo hindlimb load distribution before surgery (pre-op) and at various intervals postoperatively. Percent hindlimb load on the right (experimental) hindlimb represents a relative measure of load bearing in a quiet sitting phase of activity. Relative load distributed on right hindlimbs was decreased in groups II and 111 animals compared to preoperative and group I values. Graph shows mean values ? 1

SD.

RESULTS

ther group I1 nor group 111 animals returned to the normal preoperative values and animals continued to bear only a small fraction of their weight on their injured limb when sitting.

Evaluation of Hindlimb Immobilization

Gross Observations

The criteria for defining a knee as “immobilized” were that there was no voluntary joint movement and less than a few degrees of passive motion: with these criteria, no animal had a completely rigidly immobilized knee after surgery. Also, immobilized animals in later intervals appeared to have greater passive knee motion on manual examinations compared to early healing intervals. These observations suggest that at least some motion was present in all group 111 immobilized animals but the amount was not specifically quantified.

Knees from group I animals showed no obvious change from contralateral joint anatomy or gross tissue appearance. However, knees from group I1 animals showed gross evidence of progressive joint damage: cartilage erosions on weight-bearing surfaces, soft tissue proliferation at the joint margins, synovial thickening, and meniscal tears. In about 50% of group 11 animals, injured knees showed areas of full thickness cartilage loss on the medial femoral condyle after 6 weeks, and at 14 weeks virtually all injured joints had some full thickness cartilage loss here. The most striking finding in these knees was at the patellofemoral joint. Here, marginal proliferation of what appeared grossly to resemble fibrocartilaginous tissue was almost always present. Compared to those from group 11, group 111 knees showed less cartilage erosion of weight-bearing surfaces at 3 and 6 weeks but at 14 weeks some full thickness loss of cartilage was noted over the medial and lateral femoral condyles in about one-half of the injured knees. No meniscal pathology was found. Group 111injured knees showed some minor

“In Vivo” Hindlimb Load Distribution Animals demonstrated a ncarly even load distribution on their hindlimbs before surgery. After sham operations, group 1 animals showed virtually the same hindlimb load distributions as the preoperative values (Fig. 3). In groups I1 and 111, however, a relative decrease in right hindlimb load was found; at best, group 11 animals carried only 31% of the total load on their injured hindlimb while group 111carried only about 25%. Load distribution in nei-

J Orthop Res, Vol. 10,N o . 2 , 1992

R . C . BRAY ET AL.

162

soft tissue proliferation particularly along the medial tibial joint line, but proliferative changes were not nearly as pronounced at the patellofemoral joint compared to group I1 animals. Patellar dislocations may partially explain the gross changes seen at the patellofemoral joint. Nine animals in group I1 and four in group I11 had patellar dislocations of the injured knee at the time of sacrifice. These animals showed gross loss of articular cartilage and soft tissue proliferation at the patellofemoral joint margins. No patellar dislocations were encountered in sham-operated or contralateral control knees. Dissections of all healing complexes (groups 11 and 111) revealed evidence of MCL structural healing by recognizable “scar” tissue bridging the original 4 mm gap injury site in the MCL. Healing MCLs from group I11 were significantly smaller than those in group I1 as represented by their mean midsubstance CSA measurements (Table 1). Healing ligaments from both groups IT and HI, however, were significantly larger than shams and contralatera1 control MCL complexes at every interval tested. Biomechanical Results Laxity measurements were significantly greater in group I1 healing complexes, where values of 3 to 10 times those of both control (sham and contralaterals) and group I11 laxities were found (Table 2). The laxities of healing MCLs from group I11 animals were not significantly different from the laxities of either group I sham-operated MCLs or contralateral controls. Group IT healing complexes experienced significant decreases in laxity from 6 to 14 weeks, but still remained significantly more lax than group I11 healing complexes. In structural strength tests, both group I1 and I11 specimens had statistically lower failure forces than group I sham MCLs and contralateral controls at

every interval (Table 3). The ultimate failure force of group I1 healing MCLs improved significantly with time, becoming significantly larger than group I11 healing MCLs at 14 weeks. Force-deformation plots at 3 and 14 weeks after injury are shown in Fig. 4A and 4B, respectively. Chord stiffnesses (slopes of these plots computed between 1 .O and 2.5 mm deformation) for both group I1 and I11 healing ligament complexes were significantly lower than the stiffnesses of group I sham-operated MCLs and contralateral controls at each interval tested. At both 3 and 14 week healing intervals, the chord stiffnesses of group I11 MCLs were lower than those of group 11 MCLs, but the differences were not statistically significant. It was also apparent that the elongation to failure was lower in the immobilized group than in the nonimmobilized group at both 3 and 14 weeks of healing. Most MCL complexes failed in the midsubstance region (Table 3). Usually, failure in healing complexes occurred near the tibial junction of the scar and original ligament tissue (as determined by marker sutures). However, a few failures in group I1 and 111 healing MCLs were noted to occur at the tibial insertion site. Some femoral condylar fractures were noted in group I sham-operated MCLs and contralaterals, but tibial avulsions were not observed. Tensile strengths (MCL midsubstance stress at complex failure) of healing MCLs in both groups I1 and TI1 were significantly inferior to both those of group I sham-operated and contralateral MCLs (Fig. 5). At 14 weeks, group 11 healing MCLs attained about one-tenth of the ultimate tensile strength observed in contralaterals while group I11 healing complexes reached a mean of about onetwentieth of contralateral values. There were, however, no statistical differences in tensile strength between group I1 and 111complexes at any healing interval.

TABLE 1. Ligament midsubstance cross-sectional areas (mm2) Study interval (weeks)

Contralateral MCL complexes“

Sham MCL (group I)

3 6 14

4.0 2 1.2 3.7 f 0.5 4.1 ? 0.7

4.8 2 1.4 4.0 2 0.4 5.0 2 2.6

MCLiACL mobile (group 11)

11.92 3.1h 15.0 t 4.1b 16.7 2 3.3”

MCLiACL immobile (group 111)

7.4 t 4.3h 6.7 t 3.0b 11.4 2 3.1h

Healing MCL complexes (groups TI and 111) were significantly larger than contralateral MCL complexes at all intervals. Numbers are mean values i 1 SD. a Contralaterals from all three groups pooled at each interval. Indicates significant difference between groups I1 and IT1 (p < 0.05).

J Orthop Res, Val. 10, No. 2,1992

163

JOINT IMMOBILIZATION AND MCL HEALING TABLE 2. Ligament complex laxity (mm) Study intt:rval (weeks)

Sham MCL (group I)

Contralateral MCL complexes”

3 6 14

0.3 0.5 0.4

-C ? ?

0.2 0.3 0.2

0.4

0.4

?

0.5

* 0.1

0.2 f 0.1

MCWACL mobile (group 11)

MCLiACL immobile (group 111)

2.5 f 3.1’ 4.0 f 2.1’ 1.7 f 0.4b

0.3 ? 0.6 0.3 2 0.2 0.7 i 0.7

Numbers are mean values ? 1 SD. Contralaterals from all three groups pooled at each interval. Group I1 significantly different from all other groups at each interval (p < 0.05).

bilized joints showed less gross evidence of damage, particularly at the earliest healing intervals (3 and 6 weeks). ‘However, immobilization inhibited the formation of new MCL scar mass. Since scar material properties were similar with or without joint immobilization, the failure loads of the smaller immobilized scars were always significantly lower than those of nonimmobilized MCLs. Immobilization therefore inhibited maturation in structural properties of rabbit MCL scars in a manner similar to that noted in other healing models (12,13,21). On the positive side, some important healing MCL properties appeared to be closer to contralatera1 values in immobilized rabbits. Immobilized healing MCL laxity approached control values even after 3 weeks and remained close to control values at the later intervals tested. Thus, it appears that MCL laxity was actually improved by immobilization. Nonimmobilized MCLs probably experience more gross movement and possibly higher loads during healing than their immobilized counterparts. In an environment of increased force and displacement, nonimmobilized healing MCL scars may adapt by increasing intrinsic elongation limits, which is also evident by the increases observed in elongation to failure in nonimmobilized MCLs (Fig. 4).

DISCUSSION The results of this study generally concur with those of previous models of ACL-deficient MCL healing (3,7,8,22) and support the clinical view that ACL function is critical to MCL healing (10,19). In preliminary studies (3), we compared group I1 with a model of isolated MCL healing (4) (ACL intact, nonimmobilized joints) and showed that healing MCLs in both groups reached about the same failure loads by 14 weeks regardless of the presence or absence of the ACL. However, ACL disruption was associated with high MCL scar laxity and low MCL midsubstance stress at complex failure. ACLdeficient joints also demonstrated much more instability and joint damage in the early weeks after injury. Since those results indicated that ACLdeficient MCL healing was extremely poor, in this study we wanted to determine whether minimizing joint instability through immobilization would prevent both abnormal healing ligament laxity and progressive joint damage, without compromising the relatively good MCL scar structural failure load observed previously. The comparisons of MCL healing in ACLdeficient rabbit knees, with or without immobilization, have yielded several clear differences. Immo-

TABLE 3 . Ligament complex failure mode and failure force Study interval (weeks) 3 6 14

Contralateral MCL complexes” 266 f 77 (n = 15) 280 f 67 (n = 16) 312 f 55 (n = 15)

(M (M

(M

= = =

12, F = 3) 13, F = 3) 11,

F

MCL/ACL mobile

Sham MCL (grow 1)

=

4)

307 2 81 (n = 3) 304 f 68 (n = 3) 2 6 6 k 68 (n = 3)

MCWACL immobile (group 111)

(group 11) 1)

64 2 36

(M = 5, T

(M = 2, F = 1)

( n = 6) 54 f 19

(M

=

5, T = 2)

(M

=

6)

(M

(M

=

2, F

= 3)

=

( n = 7) 124 2 55’ (n = 6)

=

1)

22

f7 (n = 6) 37 ? 18 (n = 6) 40 ? 20’ (n = 6)

(M = 6) (M

=

5, T = 1)

(M

=

6)

M = midsubstance disruption; F = femoral fracture; T - tibia1 avulsion. Numbers are mean values (in Newtons) ? 1 SD. a Contralaterals from all three groups pooled at each interval. Groups I1 and I11 (healing MCL complexes) were significantly lower than shams and contralaterals at all healing intervals. Indicates significant difference between groups I1 and I11 (p < 0.05).

’

J Orthop Res, Vol. 10, No. 2 , 1992

164

R . C . BRAY ET A L .

In a study using a canine model of combined MCL/ACL injury, Woo and co-workers obtained different findings to those noted in the present study with respect to healing MCL load at failure. In Woo’s study, healing MCLs reached 80% of the contralateral control load to failure values, while in the present study. group IT healing MCLs obtained only 40% of contralateral values. These findings may reflect either species- or activity-specific differences between dogs and rabbits. On the other hand, healing canine MCLs from ACL-deficient knees generally showed similar cross-sectional area increases and low failure stresses compared with the rabbit results from this study. Data from in vivo limb loading tests indicate that sham-operated control animals did not alter their A

400.0

h

t

200.0

-

GROUPIII

GROUP II NOT SIG DIF GROUP 111 AT ANY DEFORMATION pz0.05

K

? 100.0

0.0 0.0

1.o

2.0

DEFORMATION

3.0

4.0

(mm)

400.0

-

+ GROUPIII

--z

; F

100.0

1n I

- -. .

GROiiP i -. .

I’ GROUPIII

GRP II NOT SIG DIF GRP 1111 AT ANY INTERVAL p>0.05 CONTRALATERALS

-

T

80.0

60.0

40.0

20.0

0.0 ..

3 WK

6 WK

14 WK

HEALING INTERVAL FIG. 5. MCL complex stress at failure as determined by dividing individual MCL failure force by their respective midsubstance cross-sectional area and pooling values for each group. Group II values were not significantly different than group 111 at any interval. Graph shows mean values 2 1 SD.

~

U

;

120.0

300.0

CONTRALATERALS

GRP II NOT SIG DIF GRP 111 AT ANY DEFORMATION

p>o.os

200.0

K

100.0

0.0

0

1.0

,

2.0

3.0

4

I

DEFORMATION (mm)

FIG. 4. Composite force-deformation plots for groups I1 and 111 and contralateral control MCL complexes at 3 (A) and 14 weeks (6) after surgery. Healing MCL complexes were weaker than controls at both intervals and showed lower structural stiffnesses as well. Graph shows mean values 5 1

SD.

Orthop Res. Vol. 10, N o , 2 , 1992

relative limb load distribution. A preliminary study on animals with a disrupted MCL but with an intact ACL also demonstrated no change in load distribution over an even longer term (40 weeks). However, in the present study, combined MCL/ACL knee injury resulted in a relative underloading of the injured limb and concomitant overloading of the contralateral limb regardless of whether immobilization was instituted (Fig. 3 ) . An interesting aspect of these tests is that injured animals in group 11 showed evidence of progressive joint damage even though the limb appeared to be bearing less load during sitting than control limbs during the early healing periods. The relative increase in hindlimb loading observed in contralateral limbs is also interesting in view of changes observed in contralateral MCL complex strength and midsubstance stress at complex failure. These contralateral MCL properties appear to improve subtly with time after opposite knee MCL injury (Table 3 and Fig. 5). In view of the known beneficial effects of exercise (and presumably increased loading) on ligaments (17,21), some of this “contralateral effect” may be mediated by increased loads on the contralateral MCL complexes. However, these speculations are somewhat limited in that our measurement only reflected static and “relative” whole limb load distributions. More dynamic and better measures of MCL load or strain environment (i.e., during animal gait) should be made to investigate these speculations specifically. There are some significant limitations to this

JOINT IMMOBILIZATION AND MCL HEALING

study. Primarily, our results are probably limited to healing MCLs in the rabbit species tested, and therefore should not be generalized. The joint immobilization technique used did not lead to absolutely rigid immobilization, and a small amount of motion was present in most immobilized knees. Perhaps another important limitation in terms of structural testing was in the dissection of healing MCL complexes, where distinguishing the borders of “healing scar” was very difficult and somewhat arbitrary. This problem has been noted in other models of combined MCL/ACL injury (3,7,22), but we agree with Woo and co-workers (22) that it is important to accept this potential error, rather than exclude any of the healing tissue mass from biomechanical tests. The assessment of “joint damage” was made with gross anatomic criteria: the actual histological and biochemical aspects of tissue damage remain unknown. Finally, the in vivo hindlimb loading tests represent only a small fraction of the load-bearing history in experimental animals. Bearing these limitations in mind, some obvious similarities and differences between immobilized and nonimmobilized animals deserve comment. For example, both groups were similar in that healing MCL complexes showed increasing scar mass, which consistently showed low ultimate force at failure and tensile strength compared to contralatera1 values. Since both group 11 and 111 animals placed less load on their injured limbs, it could be speculated that stress deprivation might have adversely affected scar remodeling and mechanical maturation, resulting in weaker healed structures. It could also be argued that immobilization reduced abnormal (and destructive) mechanical displacements of healing MCLs, leading to an additional strain or displacement deprivation state, causing proportionately less new scar material to be formed in immobilized animals (Table 1). We would conclude from this study, therefore, that immobilizing an unstable joint inhibits a gain in structural strength in healing MCL scars. However, brief immobilization ( 4 weeks or less) appeared to protect the rabbit knee joint from gross damage and produced MCL laxities closer to contralateral values than nonimmobilized healing MCLs. Immobilization can therefore be said to be good for the joint and for low-load ligament healing behavior but bad for recovery of strength and higher-load stiffnesses of this injured ligament. Since these results are short term and the effects of remobilization were not examined, longer-term

165

assessments will be necessary to define the advantages and disadvantages of immobilization in this model more fully. Acknowledgment: Financial support f o r this w o r k w a s obtained from the Alberta Heritage Foundation for Medical R e s e a r c h ( A H F M R ) , the Canadian Orthopaedic Foundation (CoreiAcore), and the Arthritis Society of Canada. We also acknowledge the expert technical assistance of Brett Hennenfent and Linda Anscomb in the completion of experiments, and L a u d McDonald for completion of t h e manuscript.

REFERENCES 1. Akeson WH, Woo SL-Y, Amiel D, Frank CB: The biology

2.

3.

4.

5. 6.

7. 8.

9. 10.

11.

12.

13.

14.

IS.

16.

of ligaments. In: Rehabilitation of the Injured Knee, ed by F Funk, LY Hunter, St. Louis, C. V. Mosby, 1984, pp 93-148 Ballock TR, Woo SL-Y, Lyon RM, Hollis MJ. Akeson WH: Use of patellar tendon autograft for anterior cruciate ligament reconstruction in the rabbit: A long-term histologic and biomechanical study. J Orthop R e s 7:474485, 1989 Bray RC, Frank C, Shrive N , Hennenfent B, Chimich D: Biomechanics of stable and unstable healing MCL complexes. Orthop Trans 13:674, 1989 Chimich D, Frank C, Shrive N , Dougall H, Bray R: The effects of initial end contact on medial collateral ligament healing-a morphological and biomechanical study in a rabbit model. J Orthop R e s 9:3747, 1991 Dixon WJ: BMDP Statistical Software. Los Angeles, University of California Press, 1983 Fetto JF, Marshall JL: The natural history and diagnosis of anterior cruciate insufficiency. Clin Orthop 147:29-38, 1980 Forbes I, Frank C, Lam T, Shrive N: The biomechanical effects of combined ligament injuries on the medial collateral ligament. Trans Orthop R e s Soc 13:186, 1988 Hart DP, Dahners LE: Healing of the medial collateral ligament in rats. The effects of repair, motion, and secondary stabilizing ligaments. J Bone Joint Surg [Am] 69:1194-1199, 1987 Holden DL, Eggert AW, Butler JE: The non-operative treatment of grade I and I1 medial collateral ligament injuries to the knee. Am J Sports Med 11:340-344, 1983 Kannus P: Long-term results of conservatively treated medial collateral ligament injuries of the knee joint. C/in Orthup 226:103-112, 1988 Noyes FR, Mooar PA, Matthews DS, Butler DL: The symptomatic anterior cruciate-deficient knee. Part I. The longterm functional disability in athletically active individuals. J Bone Joint Surg [Am] 65:154-162, 1983 Ogata K , Whiteside LA, Andersen DA: The intra-articular effect of various postoperative managements following knee ligament repair: An experimental study in dogs. Clin Orthop 1501271-276, 1980 Piper TL, Whiteside LA: Early mobilization after knee ligament repair in dogs. Clin Orthop 150:277-282, 1980 Richman RM, Barnes KO: Acute instability of the ligaments of the knee as a result of injuries to parachutists. J Bone Joint Surg 28:473490, 1946 Sandberg R, Balkfors B, Nilsson B, Westlin N: Operative versus non-operative treatment of recent injuries to the ligaments of the knee. A prospective randomized study. JBone Joint Surg [Am] 69:112G1126, 1987 Thaxter TH, Mann RA, Anderson CE: Degeneration of immobilized knee joints in rats. JBone Joint Surg [Am]47567585. 1965

J Orthop

Res, Vol. 10, No. 2 , 1992

166

R . C . BRAY ET AL.

17. Vailas AC, Tipton CM, Matthes RD, Gart M: Physical activity and its influence on the repair process of medial collateral ligaments. Connect Tissue Res 9:25-31, 1981 18. Walsh S: Immobilization affects growing ligaments. Master’s Thesis, The University of Calgary, 1989 19. Warren RF, Marshall JL: Injuries of the anterior cruciate and medial collateral ligaments of the knee. A long-term follow-up of 86 cases-part 11. Clin Orthop 136:198-211, 1978 20. Woo SL-Y, Akeson WH, Jemmott GF: Measurement of

J Orthop Res, Vol. 10, No. 2 , 1992

nonhomogenous, directional mechanical properties of articular cartilage in tension. J Biomech 9:785-791, 1976 21. Woo SL-Y, Inoue M, McGurk-Burleson E , Gomez M: Treatment of the medial collateral ligament injury. 11: Structure and function of canine knees in response to differing treatment regimens. Am J Sports Med 15:22-29, 1987 22. Woo SL-Y, Young EP, Ohland KJ, Marcin JP, Horibe S, Lin H-C: The effects of transection of the anterior cruciate ligament on healing of the medial collateral ligament. JBone Joint Surg [Am] 72:382-392, 1990