- Email: [email protected]
Maximal contraction lessens impact response in a muscle contusion model
Pergamon
MAXIMAL
PII: SOO21-9290(96)00047-4
CONTRACTION MUSCLE
J. Biomechanics, Vol. 29, No. 10, pp. 1291-1296, 1996 Copyright 0 1996 Elsewr Science Ltd. All nghts reserved Printed in Great Britain W&9290/96 $15.00 + .@I
LESSENS IMPACT RESPONSE CONTUSION MODEL
IN A
J. J. Crisco, K. D. Hentel, W. 0. Jackson, K. Goehner and P. Jokl Department of Orthopaedics and Rehabilitation, Yale University School of Medicine, New Haven, CT 06520-1701, U.S.A. Abstract-The effect of muscle contraction on a contusion injury model was studied in the gastrocnemius muscle of anesthetized rats. Both limbs of 18 rats received a contusion injury with a blunt non-penetrating impact. One hind limb was relaxed during impact and the other was electrically stimulated to tetanic contraction. The impact was produced using a drop-mass technique (mass = 171 g, height =lOl cm, spherical radius of impactor tip =6.4 mm). The impact response was determined by sampling (10 kHz) the transmitted impact force and the displacement of the impactor. In a subgroup of nine rats, the severity of the contusion injury was measured by recording contractile tension in twitch and tetanus within two hours of injury. We found that the peak impact force was significantly less (p
INTRODUCTION
Skeletal musclecontusion injuries occur frequently in sports,and have beenreported to be the most common injury in outdoor soccer(Jarvinen, 1976;Kibler, 1993; Maehlum and Daljord, 1984)and in men’sice hockey (NCAA, 1993).In other sports,contpsionstypically rank secondon averagein frequency, behind musclestrains and/or ligamentsprains(Lindenfeldet al., 1994;NCAA, 1993).Such contusioninjuriesare typically the result of a blunt, non-penetrating impact between muscleand a rigid object. The impact responseof two bodiescanbedescribedby the time courseof the impact force and displacement. These impact response variables are dependent upon numerousfactorsthat includethe body’srespectivemass, velocity, shape,and material properties.Assumingmass, velocity and shapeare constant, elastic impact theory predicts that bodiesof stiffer material (higher Young’s modulus) produce greater impact forces and greater stresses within the bodies(Greszczuk,1982;Timoshenko and Goodier, 1970). In muscle, stiffnessincreaseswith contraction. Increasesin stiffnessin the axial direction of muscle,its principle direction of action, are well appreciatedand documented (e.g., Hoffer and Andreassen,1981). In-
Received in final form 16 February 1996. Address correspondence to: J. J. Trey Crisco, Orthopaedic Research SWP-3, Rhode Island Hospital, 593 Eddy Street, Providence, RI 02903, U.S.A.
creasesin musclestiffnesswith contraction in the transversedirection of muscleare not well documented,but are readily appreciatedwith palpation.Sincemusclestiffnessis higherwhenit iscontracted,wehypothesizedthat contraction would increasethe response(e.g. peak impact force) to a blunt non-penetratingblow when compared to the relaxed muscle.Furthermore, if we assume that greater stressesare associatedwith greater injury, then we would further hypothesizethat the likelihood and the severity of contusioninjury would be greaterin contractedmusclethan in relaxedmuscle.Therefore,the purposeof this study was to determineif and how the impact responseand the severity of the impact injury differ in contractedand relaxedmuscleusinga rat model. MATERIALS
AND
METHODS
Both hind limbs of 18 male Wistar rats (240420g, Charles River Company, Kingston, NY) were impacted in the mid-belly of the gastrocnemius muscle(all experimentalprocedureswere approved by the Yale Animal Care and Use Committee).The right and left limbs were randomized,and, one limb wasimpactedwhile the gastrocnemius musclecomplex (gastrocnemius,soleus,and planteris)wasrelaxedand the other limb wasimpactedwhile the gastrocnemius complexwas stimulated to isometric tetanus. The forcedisplacement-time behavior of each impact was recorded (n = 18/group). Within two hours after impact, injury severity in a subgroup of animalswas determinedby measuringthe contractile function of the gastrocnemius
1291
1292
J. J. Crisco et a/.
complexin both limbs(n = 9/group).The other ninerats wereassignedto a survival study not reported here. Prior to impact, all rats were anesthetizedsubcutaneouslywith 0.20cc Ketamine (100gml-‘) and then intramuscularly with Ketamine and 0.30cc Xylazine (20gml-I). The shaven,posterior surface of the hind limb wasplacedup by extendingthe kneeand dorsiflexing the foot to 90”, and was then held in this isometric position with elastic bands. The limbs assignedto be contracted during impact had subdermalelectrodesaffixed using Steri-Strip (3M Surgical Division, St. Paul, MN). The musclewasstimulated(70V, 70 Hz, 0.05ms usinga SS88GrassStimulator, GrassInstrument Company, Quincy MA) to maintain contraction throughout the impact for a total duration of one second.The increasedthicknessof the limb due to musclecontraction wasmeasuredin the anterior-posterior direction at the site of impact using a standardruler with a right-angle indicator. Impacts to the mid-belly of the posterior surfaceof the gastrocnemiusmusclecomplex were produced by droppinga mass(171g)from a heightof 102cm.The apparatus was a two-masssystem,with the drop-mass contacting an impactor that wasin direct contact with the skin (Fig. 1).The impactor consistedof an aluminum cylinder (11cm long, 8 mm mid-diam,36 g) fitted on top with a sheetof rubber (1 mm thick) to reducehigh frequencyvibrations arisingfrom the drop-massandon the bottom with a nylon plastic sphere(radius=6.4 mm). This impactor designwasg modification of our previous design(Criscoet al., 1994).The impactor displacements were measuredwith a displacementtransducer(Model 1000DC-E LVDT, Schaevitz, Pennsauken,NJ), whose core was connected to the impactor by a horizontal aluminum arm (12x 9 x 56mm). The impact force required to support the limb during impact (i.e. the force transmittedthrough the limb) wasmeasuredwith a load cell (PCB Piezotronics, Depew, NY) that was secured betweenthe laboratory floor and the plastic platform that supportedthe target limb. During impact, displacement and force signalswere sampledevery 0.1 ms by computer (Macintosh Quadra 700, Apple Computer) with codedevelopedin Labview (National Instruments, Austin, TX). The load-displacement-timedata of the impactswere reduced and then compared using several impact responsevariables.Peakimpact force(N) and peak impact displacement(mm) were defined as the maximum respectivevalues recorded. The initial impact stiffness (Nmm-‘) wasestimatedusinga least-squares fit of the load-displacementcurve from 0 to 25% of the peakload. The time betweenthe peakforce and peak displacement wasdefinedas the delay (ms).The peak impact velocity (ms-l) wasthe maximumvalue of the derivative of the displacement-timecurve. Impulse (Ns) of the impact force wasthe integral of the load-time curve.The impact energy (J) was defined as the integral of the loaddisplacementcurve. Contractile function wasmeasuredin the subgroupof nine rats two hourspost-impact,while the rats werestill
Fig. 1. A schematic of the impact apparatus. The impact was produced by releasing a mass (17 1 g) through a guide tube from a height of 101 cm. The drop-mass contacted an impactor with a spherical tip (radius = 6.4 mm) which delivered a blow to the mid-belly of the gastrocnemius complex. The transmitted impact force and the displacement of the impactor were recorded by computer at 0.1 ms intervals.
under anesthesia.Further dosesof anesthesiawere administeredas neededthroughout the remainderof the procedure. The drugs usedfor anesthesiaact on the centralnervoussystemanddo not have any known effect on peripheralneuromuscularfunction. A Kirschnerwire diameter(0.45mm)was placedtransverselythrough the distalfemur (O&m abovethe kneejoint) and securedto the testingapparatus.A tourniquet, usedto help minimize lossof blood, wasplacedabout the anterior muscles only and excluded the gastrocnemiusmusclecomplex and its neurovascularsupply. Stainlesssteelsuturewire (22 gauge)waspassedthrough a drill hole in the disarticulated calcaneusand attached to a load cell (Entran Devices,Fairfield, NJ) by a smallturn-buckle, which was usedto adjustmusclelength.The sciaticnerve, bathedin paraffin oil throughout the experiment,wassecuredto an electrodethat waspassedthrough a smallskin incision. The rat’s core temperature was maintained between 98-100°F using a heat lamp. The musclecomplex was then stimulated(40v, 0.05ms duration) at 0.025in. incrementsof musclelength. The length with the greatest twitch tensionwasdefinedasthe optimal length,and this peak twitch tension was recorded.Tetanic tension was then determinedat optimal length(120V, 0.05ms,70 Hz stimulusfor 1 s). The significanceof the differencesbetweenthe impact responsevariablesand betweenthe contractile functions were determinedseparatelyusing a paired (contracted
Muscle contusion model
1293
limb vs the contralateral relaxed limb), two-tailed, Student’s t-test. The relationship betweenvarious impact variables was examined using linear regression(SYSTAT, SPSSInc., Chicago,IL).
three impactsfailed to trigger the data acquisitionsystem, thus only the impactsof 15 rats were analyzed.In addition, the time sequence of onecontractedimpactwas in error and was not used,but the forcedisplacement data remainedvalid. In terms of the specific impact variables,contraction decreased the averagepeak impact RESULTS force that was transmitted through the limb by 36%, of In contrastto our initial hypothesis,contraction signif- whileincreasingthe averagepeakimpact displacement icantly decreasedthe impact responsedescribedby the the impactor by 53%, relative to the relaxed,contralatload-time and load-displacementcurves,while increas- era1limb. Lessforce with greater displacementimplies ing the displacement-timeresponse (Fig. 2).We note that a lessstiff system.Our definition of initial impact stiffness (seethe previoussection)indicatedthat this stiffnesswas 76% lessin the contractedlimb than in the relaxedlimb. The peak impact force consistentlyoccurred before the peakimpact displacementin all impacts,asexpectedfor a viscoelasticsystem.This delay wassmallerby 44% in the contracted limb. Similarly, the impact energy was smallerby 30% in the contractedlimb. Contraction also decreasedthe force impulseof the impactsby 11%. All differencesin thesediscrete impact responsevariables werehighly significant(p < 0.01;Table 1).We found that contraction prior to impactincreasedthe thicknessof the limb by 55% on the average(relaxedthickness= 14.3+ 1.9mm) at the site of impact. 10 1.5 20 Linear correlationsexistedbetweenthe variousimpact Time (ms) (4 responsevariables when the variables from both the contractedand relaxedlimbswerecombinedand plotted. Plotting peak impact force vs peakimpact displacement, wefound that increases in force correlatedwith decreases in displacements [Fig. 3(a)]. This correlation waslinear and significant(R2 = 0.668,p < 0.01)with one apparent outlier at 553N and 10mm. Plotting peak impact displacement vs peak velocity, we found that increases in displacement correlated significantly (R' = 0.93, .._.... p ~0.01) with increasesin velocity [Fig. 3(b)]. Plotting -5 “-~...:.:.:~~ ‘... 1 force impulsevs energy absorbed,we found that, in general,increasesin the energy absorbedcorrelatedsig.,ol, I , 1 nificantly (R2 =0.66, p ~0.01) with increasesin the force 10 15 20 Time (ma] (b) impulse[Fig. 3(c)]. However, if we ungroup the impacts in Fig. 3(c),we note that the correlation washigher for P ” ” 1 ““I ” ” 1 ” ” 1 ‘I impactson relaxed limbs than for thoseon contracted limbs. On average,contractile function (twitch and tetanus) was greater two hours after impact in the musclesthat were contracted than those that were relaxed during impact (Fig. 4). At optimal length, isometrictensionwas greater in twitch (8.9 + 2.0 N vs 7.8 & 1.5 N) and in teta-
Fig. 2. The impact response of four randomly selected pairs of impacts. (a) The impact force had a faster rise time, shorter duration, and reached higher values in the relaxed limb than in the contracted limb. (b) The displacements of the impactor were greater, with less rebound, in the contracted limb. (c) The initial impact stiffness was less in the contracted limb than in the relaxed limb.
nus (31.1+ 4.3N vs 27.1+ 4.6 N). Although contractile function wasconsistentlygreaterin the contractedlimbs (Fig. 4) only the tension in tetanus was significantly different (p < 0.03).The p value for the tensionin twitch was0.2. In 24uninjured musclesfrom a separatepopulation of rats that werecomparablein ageand weight, the tetanic tensionwas43.4If: 4.9N. Comparingthesevaluesto the presentstudy suggeststhat the injuries to both the relaxed and the contracted muscles were significant (p < 0.05,unpaired,two-tailed, Student’st-test). Grossobservationof the harvestedspecimens within two hours after injury revealed disruption of muscle
1294
J. J. Crisco et al Table 1. Average [one S.D.] impact response variables for the relaxed and contracted limbs
Relaxed Contracted
Peak force (N
Peak displacement (mm)
Stiffness (Nmm-‘)
Delay (ms)
Energy Wm)
Impulse (N s)
301[77] 193 [38]
11.4[1.4] 17.4 Cl.91
71.9 [14.5] 17.3 c11.43
2.7[0.5] 1.5 CO.63
2.0 co.31 1.4 [0.5]
0.9[0.07] 0.8 [O.OS]
Note: All differences were highly significant (p < 0.01: paired, two-tailed, Student’s t-test). 400 El
(a)
Relaxed Muscle Contracted Muscle
Peak lorea (N)
.
Twitch
.
.
2.5t i 0
0
~““““““““““‘1
1
8
T&llUS
Fig. 4. Contractile function, measured isometrically at the optimal length within two hours post-injury, was greater on average in the gastrocnemius muscle complexes that werecontracted during impact than in those that were relaxed during impact. The difference in tetanus was significant (p < 0.03). Both injuries significantly decreased tetanic tension when compared with a separate uninjured control group (43.4 f 4.9 N).
gastrocnemiuscomplexesbetweenthe contracted limb (2.5f 0.4 g) and the relaxedlimb (2.6f 0.3 g). DISCUSSION
The question addressedby the present study was whethermaximalmusclecontraction alters the response of muscleto blunt non-penetratingimpacts.Contrary to % our initial hypothesis,the impact response of the contrac0. a . ted limb was less than that of the relaxed limb. Specifi. #. cally, contraction decreasedthe impact stiffnessof the limb in the transverse direction (peak transmitted . . force decreasedwhile peak displacement increased 0.7 I, , , with respectto the relaxed limb). Contraction also de0.5 1.0 1.5 2.0 Energy absorbed (Nmm) Cc) creasedthe viscousresponseof the limb (impact energy and delay decreasedwith respectto the relaxed limb). Fig. 3. Linear correlations between various impact response Contraction did not increaseinjury severity and may variables for all impacts were examined by plotting (a) peak have acted as a mechanismto reduce the severity of impact force vs peak impact displacement, (b) peak impact displacement vs peak energy, and (c) impact energy vs force a contusioninjury. The potential protective mechanism impulse. of contraction was suggestedby the ability of these musclesto generatehigher tetanic tensions.Since we tissue at the site of impact with intramuscular-interstitial were not able to document a significant difference hematoma.We could discern no grossdifferencesbe- betweenthe twitch tensions,the degreeto which contractweenthe musclesof the contractedlimb and thoseof the tion acts as a protective mechanismremains to be. relaxedlimb. There wasno differencein the massof the determined.
Muscle contusion model The documented affect of muscle contraction on impact response may be explained by considering the limb as a composite of both muscle and underlying bone. Maximal muscle contraction decreased the impact stiffness of the limb most likely because the contracted muscle reduced the influence of the underlying bone. Muscle tissue alone is stiffer when it is contracted then when it is relaxed, both in the longitudinal direction (Hoffer and Andreassen, 1981) and in the transverse direction. If we consider the impact of a composite of muscle and bone, our data suggests that, in the relaxed limb the impactor tip at rest was in close approximation to the underlying bone. At the initiation of loading, the ’ impactor readily compressed the remaining muscle tissue against the bone and the impact force readily increased with little increase in the displacement. These increases in impact force with small increases in displacement define a relatively stiff composite. With contraction, the stiffness of the muscle tissue, that is its resistance to displacement by the impactor, and the thickness of the mid muscle belly, which is the site of impact, both increased. This increased stiffness and thickness of the muscle during impact combined to reduce the influence of the underlying bone while permitting greater displacements of the impactor [Fig. 3(a)]. This effectively resulted in a limb that was less stiff than the composite of relaxed muscle and bone. It must be noted that these findings pertain to the whole limb and are most likely dependent upon numerous factors including muscle morphology, level of activation, and underlying bony geometry. Experimental muscle contusion injuries have been produced previously using a spring-loaded hammer (Jarvinen, 1975, 1976a, b; Jarvinen and Sorvari, 1975) and a drop mass technique (Fisher et al., 1990; Stauber et al., 1990;Stratton et al., 1984).Thesestudiesand our previous study (Criscoet al., 1994)have demonstratedthat consistentimpactsproducea consistentcontusioninjury. One advantageof our modelis the ability to record the forcedisplacement-time behavior of the impact response.The limitations of our modelincludethe following. The actual mechanismof injury and the thresholdof injury werenot determined.Although the impactmechanismis simple,it doesconsistof a two-masssystem[the drop-mass(171g) and the impactor (3Og)]. While this systemproduces a consistentimpact, the engineering analysisof a two-masssystemimpact is more complicated than the analysisof a single masssystem.The advantageof the two-masssystemis the easewith which the impact responsecan be measuredand the target area can be defined. Hence,our model may be a closerapproximation of an impact to a pieceof (rigid) equipment that lies over the muscle,whereasa singlemasssystem may be a closerapproximation of a direct blow to the muscle.A further limitation of our present model is that wedid not measurethe impact force at the impactor tip. How the magnitudeand time courseof this force compareswith the transmitted impact force we did measureremainsto be determined.In our model we are clearly studying the composite responseof the whole muscleand the underlying bone; we are not able
1295
to differentiate muscleresponsefrom that of the compositeresponse. In relating the increasedstiffnessof the relaxedlimb to increasedinjury, we are making the fundamental assumptionthat increasedstresses are associatedwith increasedinjury. This assumptionis clearly valid if injury is associatedwith materialfailure, that is, grossdamageto tissueintegrity. Weobservedgrosstissuedamagein both impactsand both impactsproduceda significantinjury in the contractile function of the musclecomplex.However,in this study, we did not quantify the extent of tissue damage.Moreover, injury may occur to the neuromuscular componentswithout grossdamage.Which internal impact stresses (compression, tension,or shear)produce injury isa complexquestionthat remainsunclear.A similar question has been examinedin experimental and theoretical impact studiesof lung tissues(Fung et al., 1988)andbrain tissues(Chu et al., 1994).In thosestudies, tensilestresses in lung tissuesand shearstresses in brain tissuewerehypothesizedto be the principal injury mechanism. The ability of contractedmuscleto influencethe injury mechanismand to reducethe severity of injury hasbeen documentedin studiesof musclestrain injuries (Garret et al., 1987)and long bone fractures (Nordsletten and Ekland, 1993).As a model of musclestrain injury, the extensordigitorum longusmuscleof rabbitswaslongitudinally pulled to failure in tension.Paired comparisons (contracted vs relaxed) of the failure variables demonstratedthat the contractedmusclefailed at a significantly greaterforce and absorbeda significantlygreateramount of energy.In the study on long bone fractures,the fracture strengthandultimate energyabsorptionwereshown to be significantly greaterin limbswith contracted musculature,ascomparedto the contralateralrelaxedlimb in a rat model. In contrast to thesefindings and to our findings,Robinovitch et al. (1991)predictedthat contracted musclewould increasethe likelihood of hip fractures. When volunteerscontracted their hip musclesthe effective stiffnessof the body increased,producing higher impact forces, which would increasethe likelihood of femoralfracture. The discrepancyin our resultsmay be due to the differencebetweenimpactsdirectly to bone and impactsdirectly to muscle. In summary,contraction during the non-penetrating impact significantly alteredthe impact responsecreating an effectively lessstiff limb than when the musclewas relaxed during impact. The contracted limb was not more susceptiblethan the relaxedlimb to impact injury. On the contrary, our data suggeststhat contracted musclemay evenreducethe severity of contusioninjury. Thisfinding would corroboratethe protective reactionof someone‘tightening their’ musclesprior to receiving a blow in an attempt to minimizethe pain and/or injury. At present,there is limited data availableon the impact of muscleand limbs.Experimentssuchasthis may aid in the prevention of this frequentinjury by providing essential data on the behavior and the tolerance of whole muscleto impact. Suchdata may be usefulin the design of protective sportsequipment.
1296
J. J. Crisco et a/.
Acknowledgements-The authors acknowledge The Whitaker Foundation for their financial support, and John Choi and Jay Huddleston for their assistance. REFERENCES
Chu, C. S., Lin, M. S.,Huang, H. M. and Lee, M. C. (1994) Finite element analysis of cerebral contusion. J. Biomechanics 27, 187-194. Crisco, J. J., Jokl, P., Heinen, G. T., Connell, M. D. and Panjabi, M. M. (1994) A muscle contusion injury model: biomechanics, ohvsiolow. and histologv. Am. J. Sports Med. 22, 702-710. Fisher, B. D:,Baracos, V. E.; Shnitka, T. K., Mendryk, S. W. and Reid, D. C. (1990) Ultrastructural events following acute muscle trauma. Med. Sci. Sports Exer. 22(2), 185-193. Fung, Y. C., Yen, R. T., Tao, Z. L. and Liu, S. Q. (1988) A hypothesis on the mechanism of trauma of lung tissue subjected to impact load. J Biomech. Engng 110, 50-56. Garret, W., Safran, M., Seaber, A., Glisson, R. and Ribbeck, B. (1987) Biomechanical comparison of stimulated and nonstimulated skeletal muscle pulled to failure. Am. J. Sports Med.
E(5),
448-454.
Greszczuk, L. B. (1982) Damage in composite materials due to low velocity impact. In Impact Dynamics (edited by Zukas, J. A., Nicholas, T., Swift, H. F., Greszczuk, L. B. and Curran, D. R.) Wiley, New York. Hoffer, J. A. and Andreassen, S. (1981) Regulation of soleus muscle stiffness in premammillary cats: intrinsic and reflex components. J. Neirophysiol. 45, 267-285. Jarvinen. M. (1975) Healing of a crush iniurv in rat striated muscle. 2. A histological study of the e&t of early mobilization and immobilization on the repair processes. Acta Path. Microbial.
Stand.
Sect A 83, 269-282.
Jarvinen, M. (1976a) Healing of a crush injury in rat striated muscle. 3. A micro-angiographical study of the effect of early
mobilization and immobilization Path. Microbial.
Stand.
on capillary ingrowth. Acta
Sect A 84, 85-94.
Jarvinen. M. (1976b) Healing of a crush injury in rat striated muscle. 4. Effect of early mobilization and immobilization on the tensile properties of gastrocnemius muscle. Acta Chir. Stand.
142,47-56.
Jarvinen, M. and Sorvari, T. (1975) Healing of a crush injury in rat striated muscle. 1. Description and testing of a new method of inducing a standard injury to the calf muscles. Acta Path. Microbial.
Stand.
Sect A 83, 259-265.
Kibler, W. B. (1993) Injuries in adolescent and preadolescent soccer players. Med. Sci. Sports Exer. 25( 12), 1330-1332. Lindenfeld, T. N., Schmitt, D. J., Hendy, M. P., Mangine, R. E. and Noyes, F. R. (1994) Incidence of injury in indoor soccer. Am. J. Sports Med. 22(3), 364-371. Maehlum, S. and Daljord, 0. A. (1984) Football injuries in Oslo. A one year study. Br. J. Sports Med. 18, 186-190. National Collegiate Athletic Association (1989-1993) Injury Surveillance System. Overland Park, Kansas. Nordsletten, L. and Ekeland, A. (1993) Muscle contraction increasesthe structural capacity of the lower leg: an in vivo study in the rat. J. Orthop. Res. 11, 299-304. Robinovitch, S. N., Hayes, W. C. and McMahon, T. A. (1991) Prediction of femoral impact forces in falls on the hip. J. Biomech.
Engng
113, 366-374.
Stauber, W. T., Fritz, V. K. and Dahlmann, B. (1990) Extracellular matrix changes following blunt trauma to rat skeletal muscles. Exp. Mol. Pathol. 52, 69-86. Stratton, S. A., Heckman, R and Francis, R. S. (1984) Therapeutic ultrasound, its effect on the integrity of a nonpenetrating wound. J. Ortho. Sports. Phys. Ther. 5, 278-281. Sullivan, J. A. and Gross, D. H. (1980) Evaluation of injuries in youth soccer. Am. J. Sports Med. 8, 325. Timoshenko, S. P. and Goodier, J. N. (1970) Theory ofElasticity (3rd Edn) p. 420. McGraw-Hill, New York.
MAXIMAL
PII: SOO21-9290(96)00047-4
CONTRACTION MUSCLE
J. Biomechanics, Vol. 29, No. 10, pp. 1291-1296, 1996 Copyright 0 1996 Elsewr Science Ltd. All nghts reserved Printed in Great Britain W&9290/96 $15.00 + .@I
LESSENS IMPACT RESPONSE CONTUSION MODEL
IN A
J. J. Crisco, K. D. Hentel, W. 0. Jackson, K. Goehner and P. Jokl Department of Orthopaedics and Rehabilitation, Yale University School of Medicine, New Haven, CT 06520-1701, U.S.A. Abstract-The effect of muscle contraction on a contusion injury model was studied in the gastrocnemius muscle of anesthetized rats. Both limbs of 18 rats received a contusion injury with a blunt non-penetrating impact. One hind limb was relaxed during impact and the other was electrically stimulated to tetanic contraction. The impact was produced using a drop-mass technique (mass = 171 g, height =lOl cm, spherical radius of impactor tip =6.4 mm). The impact response was determined by sampling (10 kHz) the transmitted impact force and the displacement of the impactor. In a subgroup of nine rats, the severity of the contusion injury was measured by recording contractile tension in twitch and tetanus within two hours of injury. We found that the peak impact force was significantly less (p
INTRODUCTION
Skeletal musclecontusion injuries occur frequently in sports,and have beenreported to be the most common injury in outdoor soccer(Jarvinen, 1976;Kibler, 1993; Maehlum and Daljord, 1984)and in men’sice hockey (NCAA, 1993).In other sports,contpsionstypically rank secondon averagein frequency, behind musclestrains and/or ligamentsprains(Lindenfeldet al., 1994;NCAA, 1993).Such contusioninjuriesare typically the result of a blunt, non-penetrating impact between muscleand a rigid object. The impact responseof two bodiescanbedescribedby the time courseof the impact force and displacement. These impact response variables are dependent upon numerousfactorsthat includethe body’srespectivemass, velocity, shape,and material properties.Assumingmass, velocity and shapeare constant, elastic impact theory predicts that bodiesof stiffer material (higher Young’s modulus) produce greater impact forces and greater stresses within the bodies(Greszczuk,1982;Timoshenko and Goodier, 1970). In muscle, stiffnessincreaseswith contraction. Increasesin stiffnessin the axial direction of muscle,its principle direction of action, are well appreciatedand documented (e.g., Hoffer and Andreassen,1981). In-
Received in final form 16 February 1996. Address correspondence to: J. J. Trey Crisco, Orthopaedic Research SWP-3, Rhode Island Hospital, 593 Eddy Street, Providence, RI 02903, U.S.A.
creasesin musclestiffnesswith contraction in the transversedirection of muscleare not well documented,but are readily appreciatedwith palpation.Sincemusclestiffnessis higherwhenit iscontracted,wehypothesizedthat contraction would increasethe response(e.g. peak impact force) to a blunt non-penetratingblow when compared to the relaxed muscle.Furthermore, if we assume that greater stressesare associatedwith greater injury, then we would further hypothesizethat the likelihood and the severity of contusioninjury would be greaterin contractedmusclethan in relaxedmuscle.Therefore,the purposeof this study was to determineif and how the impact responseand the severity of the impact injury differ in contractedand relaxedmuscleusinga rat model. MATERIALS
AND
METHODS
Both hind limbs of 18 male Wistar rats (240420g, Charles River Company, Kingston, NY) were impacted in the mid-belly of the gastrocnemius muscle(all experimentalprocedureswere approved by the Yale Animal Care and Use Committee).The right and left limbs were randomized,and, one limb wasimpactedwhile the gastrocnemius musclecomplex (gastrocnemius,soleus,and planteris)wasrelaxedand the other limb wasimpactedwhile the gastrocnemius complexwas stimulated to isometric tetanus. The forcedisplacement-time behavior of each impact was recorded (n = 18/group). Within two hours after impact, injury severity in a subgroup of animalswas determinedby measuringthe contractile function of the gastrocnemius
1291
1292
J. J. Crisco et a/.
complexin both limbs(n = 9/group).The other ninerats wereassignedto a survival study not reported here. Prior to impact, all rats were anesthetizedsubcutaneouslywith 0.20cc Ketamine (100gml-‘) and then intramuscularly with Ketamine and 0.30cc Xylazine (20gml-I). The shaven,posterior surface of the hind limb wasplacedup by extendingthe kneeand dorsiflexing the foot to 90”, and was then held in this isometric position with elastic bands. The limbs assignedto be contracted during impact had subdermalelectrodesaffixed using Steri-Strip (3M Surgical Division, St. Paul, MN). The musclewasstimulated(70V, 70 Hz, 0.05ms usinga SS88GrassStimulator, GrassInstrument Company, Quincy MA) to maintain contraction throughout the impact for a total duration of one second.The increasedthicknessof the limb due to musclecontraction wasmeasuredin the anterior-posterior direction at the site of impact using a standardruler with a right-angle indicator. Impacts to the mid-belly of the posterior surfaceof the gastrocnemiusmusclecomplex were produced by droppinga mass(171g)from a heightof 102cm.The apparatus was a two-masssystem,with the drop-mass contacting an impactor that wasin direct contact with the skin (Fig. 1).The impactor consistedof an aluminum cylinder (11cm long, 8 mm mid-diam,36 g) fitted on top with a sheetof rubber (1 mm thick) to reducehigh frequencyvibrations arisingfrom the drop-massandon the bottom with a nylon plastic sphere(radius=6.4 mm). This impactor designwasg modification of our previous design(Criscoet al., 1994).The impactor displacements were measuredwith a displacementtransducer(Model 1000DC-E LVDT, Schaevitz, Pennsauken,NJ), whose core was connected to the impactor by a horizontal aluminum arm (12x 9 x 56mm). The impact force required to support the limb during impact (i.e. the force transmittedthrough the limb) wasmeasuredwith a load cell (PCB Piezotronics, Depew, NY) that was secured betweenthe laboratory floor and the plastic platform that supportedthe target limb. During impact, displacement and force signalswere sampledevery 0.1 ms by computer (Macintosh Quadra 700, Apple Computer) with codedevelopedin Labview (National Instruments, Austin, TX). The load-displacement-timedata of the impactswere reduced and then compared using several impact responsevariables.Peakimpact force(N) and peak impact displacement(mm) were defined as the maximum respectivevalues recorded. The initial impact stiffness (Nmm-‘) wasestimatedusinga least-squares fit of the load-displacementcurve from 0 to 25% of the peakload. The time betweenthe peakforce and peak displacement wasdefinedas the delay (ms).The peak impact velocity (ms-l) wasthe maximumvalue of the derivative of the displacement-timecurve. Impulse (Ns) of the impact force wasthe integral of the load-time curve.The impact energy (J) was defined as the integral of the loaddisplacementcurve. Contractile function wasmeasuredin the subgroupof nine rats two hourspost-impact,while the rats werestill
Fig. 1. A schematic of the impact apparatus. The impact was produced by releasing a mass (17 1 g) through a guide tube from a height of 101 cm. The drop-mass contacted an impactor with a spherical tip (radius = 6.4 mm) which delivered a blow to the mid-belly of the gastrocnemius complex. The transmitted impact force and the displacement of the impactor were recorded by computer at 0.1 ms intervals.
under anesthesia.Further dosesof anesthesiawere administeredas neededthroughout the remainderof the procedure. The drugs usedfor anesthesiaact on the centralnervoussystemanddo not have any known effect on peripheralneuromuscularfunction. A Kirschnerwire diameter(0.45mm)was placedtransverselythrough the distalfemur (O&m abovethe kneejoint) and securedto the testingapparatus.A tourniquet, usedto help minimize lossof blood, wasplacedabout the anterior muscles only and excluded the gastrocnemiusmusclecomplex and its neurovascularsupply. Stainlesssteelsuturewire (22 gauge)waspassedthrough a drill hole in the disarticulated calcaneusand attached to a load cell (Entran Devices,Fairfield, NJ) by a smallturn-buckle, which was usedto adjustmusclelength.The sciaticnerve, bathedin paraffin oil throughout the experiment,wassecuredto an electrodethat waspassedthrough a smallskin incision. The rat’s core temperature was maintained between 98-100°F using a heat lamp. The musclecomplex was then stimulated(40v, 0.05ms duration) at 0.025in. incrementsof musclelength. The length with the greatest twitch tensionwasdefinedasthe optimal length,and this peak twitch tension was recorded.Tetanic tension was then determinedat optimal length(120V, 0.05ms,70 Hz stimulusfor 1 s). The significanceof the differencesbetweenthe impact responsevariablesand betweenthe contractile functions were determinedseparatelyusing a paired (contracted
Muscle contusion model
1293
limb vs the contralateral relaxed limb), two-tailed, Student’s t-test. The relationship betweenvarious impact variables was examined using linear regression(SYSTAT, SPSSInc., Chicago,IL).
three impactsfailed to trigger the data acquisitionsystem, thus only the impactsof 15 rats were analyzed.In addition, the time sequence of onecontractedimpactwas in error and was not used,but the forcedisplacement data remainedvalid. In terms of the specific impact variables,contraction decreased the averagepeak impact RESULTS force that was transmitted through the limb by 36%, of In contrastto our initial hypothesis,contraction signif- whileincreasingthe averagepeakimpact displacement icantly decreasedthe impact responsedescribedby the the impactor by 53%, relative to the relaxed,contralatload-time and load-displacementcurves,while increas- era1limb. Lessforce with greater displacementimplies ing the displacement-timeresponse (Fig. 2).We note that a lessstiff system.Our definition of initial impact stiffness (seethe previoussection)indicatedthat this stiffnesswas 76% lessin the contractedlimb than in the relaxedlimb. The peak impact force consistentlyoccurred before the peakimpact displacementin all impacts,asexpectedfor a viscoelasticsystem.This delay wassmallerby 44% in the contracted limb. Similarly, the impact energy was smallerby 30% in the contractedlimb. Contraction also decreasedthe force impulseof the impactsby 11%. All differencesin thesediscrete impact responsevariables werehighly significant(p < 0.01;Table 1).We found that contraction prior to impactincreasedthe thicknessof the limb by 55% on the average(relaxedthickness= 14.3+ 1.9mm) at the site of impact. 10 1.5 20 Linear correlationsexistedbetweenthe variousimpact Time (ms) (4 responsevariables when the variables from both the contractedand relaxedlimbswerecombinedand plotted. Plotting peak impact force vs peakimpact displacement, wefound that increases in force correlatedwith decreases in displacements [Fig. 3(a)]. This correlation waslinear and significant(R2 = 0.668,p < 0.01)with one apparent outlier at 553N and 10mm. Plotting peak impact displacement vs peak velocity, we found that increases in displacement correlated significantly (R' = 0.93, .._.... p ~0.01) with increasesin velocity [Fig. 3(b)]. Plotting -5 “-~...:.:.:~~ ‘... 1 force impulsevs energy absorbed,we found that, in general,increasesin the energy absorbedcorrelatedsig.,ol, I , 1 nificantly (R2 =0.66, p ~0.01) with increasesin the force 10 15 20 Time (ma] (b) impulse[Fig. 3(c)]. However, if we ungroup the impacts in Fig. 3(c),we note that the correlation washigher for P ” ” 1 ““I ” ” 1 ” ” 1 ‘I impactson relaxed limbs than for thoseon contracted limbs. On average,contractile function (twitch and tetanus) was greater two hours after impact in the musclesthat were contracted than those that were relaxed during impact (Fig. 4). At optimal length, isometrictensionwas greater in twitch (8.9 + 2.0 N vs 7.8 & 1.5 N) and in teta-
Fig. 2. The impact response of four randomly selected pairs of impacts. (a) The impact force had a faster rise time, shorter duration, and reached higher values in the relaxed limb than in the contracted limb. (b) The displacements of the impactor were greater, with less rebound, in the contracted limb. (c) The initial impact stiffness was less in the contracted limb than in the relaxed limb.
nus (31.1+ 4.3N vs 27.1+ 4.6 N). Although contractile function wasconsistentlygreaterin the contractedlimbs (Fig. 4) only the tension in tetanus was significantly different (p < 0.03).The p value for the tensionin twitch was0.2. In 24uninjured musclesfrom a separatepopulation of rats that werecomparablein ageand weight, the tetanic tensionwas43.4If: 4.9N. Comparingthesevaluesto the presentstudy suggeststhat the injuries to both the relaxed and the contracted muscles were significant (p < 0.05,unpaired,two-tailed, Student’st-test). Grossobservationof the harvestedspecimens within two hours after injury revealed disruption of muscle
1294
J. J. Crisco et al Table 1. Average [one S.D.] impact response variables for the relaxed and contracted limbs
Relaxed Contracted
Peak force (N
Peak displacement (mm)
Stiffness (Nmm-‘)
Delay (ms)
Energy Wm)
Impulse (N s)
301[77] 193 [38]
11.4[1.4] 17.4 Cl.91
71.9 [14.5] 17.3 c11.43
2.7[0.5] 1.5 CO.63
2.0 co.31 1.4 [0.5]
0.9[0.07] 0.8 [O.OS]
Note: All differences were highly significant (p < 0.01: paired, two-tailed, Student’s t-test). 400 El
(a)
Relaxed Muscle Contracted Muscle
Peak lorea (N)
.
Twitch
.
.
2.5t i 0
0
~““““““““““‘1
1
8
T&llUS
Fig. 4. Contractile function, measured isometrically at the optimal length within two hours post-injury, was greater on average in the gastrocnemius muscle complexes that werecontracted during impact than in those that were relaxed during impact. The difference in tetanus was significant (p < 0.03). Both injuries significantly decreased tetanic tension when compared with a separate uninjured control group (43.4 f 4.9 N).
gastrocnemiuscomplexesbetweenthe contracted limb (2.5f 0.4 g) and the relaxedlimb (2.6f 0.3 g). DISCUSSION
The question addressedby the present study was whethermaximalmusclecontraction alters the response of muscleto blunt non-penetratingimpacts.Contrary to % our initial hypothesis,the impact response of the contrac0. a . ted limb was less than that of the relaxed limb. Specifi. #. cally, contraction decreasedthe impact stiffnessof the limb in the transverse direction (peak transmitted . . force decreasedwhile peak displacement increased 0.7 I, , , with respectto the relaxed limb). Contraction also de0.5 1.0 1.5 2.0 Energy absorbed (Nmm) Cc) creasedthe viscousresponseof the limb (impact energy and delay decreasedwith respectto the relaxed limb). Fig. 3. Linear correlations between various impact response Contraction did not increaseinjury severity and may variables for all impacts were examined by plotting (a) peak have acted as a mechanismto reduce the severity of impact force vs peak impact displacement, (b) peak impact displacement vs peak energy, and (c) impact energy vs force a contusioninjury. The potential protective mechanism impulse. of contraction was suggestedby the ability of these musclesto generatehigher tetanic tensions.Since we tissue at the site of impact with intramuscular-interstitial were not able to document a significant difference hematoma.We could discern no grossdifferencesbe- betweenthe twitch tensions,the degreeto which contractweenthe musclesof the contractedlimb and thoseof the tion acts as a protective mechanismremains to be. relaxedlimb. There wasno differencein the massof the determined.
Muscle contusion model The documented affect of muscle contraction on impact response may be explained by considering the limb as a composite of both muscle and underlying bone. Maximal muscle contraction decreased the impact stiffness of the limb most likely because the contracted muscle reduced the influence of the underlying bone. Muscle tissue alone is stiffer when it is contracted then when it is relaxed, both in the longitudinal direction (Hoffer and Andreassen, 1981) and in the transverse direction. If we consider the impact of a composite of muscle and bone, our data suggests that, in the relaxed limb the impactor tip at rest was in close approximation to the underlying bone. At the initiation of loading, the ’ impactor readily compressed the remaining muscle tissue against the bone and the impact force readily increased with little increase in the displacement. These increases in impact force with small increases in displacement define a relatively stiff composite. With contraction, the stiffness of the muscle tissue, that is its resistance to displacement by the impactor, and the thickness of the mid muscle belly, which is the site of impact, both increased. This increased stiffness and thickness of the muscle during impact combined to reduce the influence of the underlying bone while permitting greater displacements of the impactor [Fig. 3(a)]. This effectively resulted in a limb that was less stiff than the composite of relaxed muscle and bone. It must be noted that these findings pertain to the whole limb and are most likely dependent upon numerous factors including muscle morphology, level of activation, and underlying bony geometry. Experimental muscle contusion injuries have been produced previously using a spring-loaded hammer (Jarvinen, 1975, 1976a, b; Jarvinen and Sorvari, 1975) and a drop mass technique (Fisher et al., 1990; Stauber et al., 1990;Stratton et al., 1984).Thesestudiesand our previous study (Criscoet al., 1994)have demonstratedthat consistentimpactsproducea consistentcontusioninjury. One advantageof our modelis the ability to record the forcedisplacement-time behavior of the impact response.The limitations of our modelincludethe following. The actual mechanismof injury and the thresholdof injury werenot determined.Although the impactmechanismis simple,it doesconsistof a two-masssystem[the drop-mass(171g) and the impactor (3Og)]. While this systemproduces a consistentimpact, the engineering analysisof a two-masssystemimpact is more complicated than the analysisof a single masssystem.The advantageof the two-masssystemis the easewith which the impact responsecan be measuredand the target area can be defined. Hence,our model may be a closerapproximation of an impact to a pieceof (rigid) equipment that lies over the muscle,whereasa singlemasssystem may be a closerapproximation of a direct blow to the muscle.A further limitation of our present model is that wedid not measurethe impact force at the impactor tip. How the magnitudeand time courseof this force compareswith the transmitted impact force we did measureremainsto be determined.In our model we are clearly studying the composite responseof the whole muscleand the underlying bone; we are not able
1295
to differentiate muscleresponsefrom that of the compositeresponse. In relating the increasedstiffnessof the relaxedlimb to increasedinjury, we are making the fundamental assumptionthat increasedstresses are associatedwith increasedinjury. This assumptionis clearly valid if injury is associatedwith materialfailure, that is, grossdamageto tissueintegrity. Weobservedgrosstissuedamagein both impactsand both impactsproduceda significantinjury in the contractile function of the musclecomplex.However,in this study, we did not quantify the extent of tissue damage.Moreover, injury may occur to the neuromuscular componentswithout grossdamage.Which internal impact stresses (compression, tension,or shear)produce injury isa complexquestionthat remainsunclear.A similar question has been examinedin experimental and theoretical impact studiesof lung tissues(Fung et al., 1988)andbrain tissues(Chu et al., 1994).In thosestudies, tensilestresses in lung tissuesand shearstresses in brain tissuewerehypothesizedto be the principal injury mechanism. The ability of contractedmuscleto influencethe injury mechanismand to reducethe severity of injury hasbeen documentedin studiesof musclestrain injuries (Garret et al., 1987)and long bone fractures (Nordsletten and Ekland, 1993).As a model of musclestrain injury, the extensordigitorum longusmuscleof rabbitswaslongitudinally pulled to failure in tension.Paired comparisons (contracted vs relaxed) of the failure variables demonstratedthat the contractedmusclefailed at a significantly greaterforce and absorbeda significantlygreateramount of energy.In the study on long bone fractures,the fracture strengthandultimate energyabsorptionwereshown to be significantly greaterin limbswith contracted musculature,ascomparedto the contralateralrelaxedlimb in a rat model. In contrast to thesefindings and to our findings,Robinovitch et al. (1991)predictedthat contracted musclewould increasethe likelihood of hip fractures. When volunteerscontracted their hip musclesthe effective stiffnessof the body increased,producing higher impact forces, which would increasethe likelihood of femoralfracture. The discrepancyin our resultsmay be due to the differencebetweenimpactsdirectly to bone and impactsdirectly to muscle. In summary,contraction during the non-penetrating impact significantly alteredthe impact responsecreating an effectively lessstiff limb than when the musclewas relaxed during impact. The contracted limb was not more susceptiblethan the relaxedlimb to impact injury. On the contrary, our data suggeststhat contracted musclemay evenreducethe severity of contusioninjury. Thisfinding would corroboratethe protective reactionof someone‘tightening their’ musclesprior to receiving a blow in an attempt to minimizethe pain and/or injury. At present,there is limited data availableon the impact of muscleand limbs.Experimentssuchasthis may aid in the prevention of this frequentinjury by providing essential data on the behavior and the tolerance of whole muscleto impact. Suchdata may be usefulin the design of protective sportsequipment.
1296
J. J. Crisco et a/.
Acknowledgements-The authors acknowledge The Whitaker Foundation for their financial support, and John Choi and Jay Huddleston for their assistance. REFERENCES
Chu, C. S., Lin, M. S.,Huang, H. M. and Lee, M. C. (1994) Finite element analysis of cerebral contusion. J. Biomechanics 27, 187-194. Crisco, J. J., Jokl, P., Heinen, G. T., Connell, M. D. and Panjabi, M. M. (1994) A muscle contusion injury model: biomechanics, ohvsiolow. and histologv. Am. J. Sports Med. 22, 702-710. Fisher, B. D:,Baracos, V. E.; Shnitka, T. K., Mendryk, S. W. and Reid, D. C. (1990) Ultrastructural events following acute muscle trauma. Med. Sci. Sports Exer. 22(2), 185-193. Fung, Y. C., Yen, R. T., Tao, Z. L. and Liu, S. Q. (1988) A hypothesis on the mechanism of trauma of lung tissue subjected to impact load. J Biomech. Engng 110, 50-56. Garret, W., Safran, M., Seaber, A., Glisson, R. and Ribbeck, B. (1987) Biomechanical comparison of stimulated and nonstimulated skeletal muscle pulled to failure. Am. J. Sports Med.
E(5),
448-454.
Greszczuk, L. B. (1982) Damage in composite materials due to low velocity impact. In Impact Dynamics (edited by Zukas, J. A., Nicholas, T., Swift, H. F., Greszczuk, L. B. and Curran, D. R.) Wiley, New York. Hoffer, J. A. and Andreassen, S. (1981) Regulation of soleus muscle stiffness in premammillary cats: intrinsic and reflex components. J. Neirophysiol. 45, 267-285. Jarvinen. M. (1975) Healing of a crush iniurv in rat striated muscle. 2. A histological study of the e&t of early mobilization and immobilization on the repair processes. Acta Path. Microbial.
Stand.
Sect A 83, 269-282.
Jarvinen, M. (1976a) Healing of a crush injury in rat striated muscle. 3. A micro-angiographical study of the effect of early
mobilization and immobilization Path. Microbial.
Stand.
on capillary ingrowth. Acta
Sect A 84, 85-94.
Jarvinen. M. (1976b) Healing of a crush injury in rat striated muscle. 4. Effect of early mobilization and immobilization on the tensile properties of gastrocnemius muscle. Acta Chir. Stand.
142,47-56.
Jarvinen, M. and Sorvari, T. (1975) Healing of a crush injury in rat striated muscle. 1. Description and testing of a new method of inducing a standard injury to the calf muscles. Acta Path. Microbial.
Stand.
Sect A 83, 259-265.
Kibler, W. B. (1993) Injuries in adolescent and preadolescent soccer players. Med. Sci. Sports Exer. 25( 12), 1330-1332. Lindenfeld, T. N., Schmitt, D. J., Hendy, M. P., Mangine, R. E. and Noyes, F. R. (1994) Incidence of injury in indoor soccer. Am. J. Sports Med. 22(3), 364-371. Maehlum, S. and Daljord, 0. A. (1984) Football injuries in Oslo. A one year study. Br. J. Sports Med. 18, 186-190. National Collegiate Athletic Association (1989-1993) Injury Surveillance System. Overland Park, Kansas. Nordsletten, L. and Ekeland, A. (1993) Muscle contraction increasesthe structural capacity of the lower leg: an in vivo study in the rat. J. Orthop. Res. 11, 299-304. Robinovitch, S. N., Hayes, W. C. and McMahon, T. A. (1991) Prediction of femoral impact forces in falls on the hip. J. Biomech.
Engng
113, 366-374.
Stauber, W. T., Fritz, V. K. and Dahlmann, B. (1990) Extracellular matrix changes following blunt trauma to rat skeletal muscles. Exp. Mol. Pathol. 52, 69-86. Stratton, S. A., Heckman, R and Francis, R. S. (1984) Therapeutic ultrasound, its effect on the integrity of a nonpenetrating wound. J. Ortho. Sports. Phys. Ther. 5, 278-281. Sullivan, J. A. and Gross, D. H. (1980) Evaluation of injuries in youth soccer. Am. J. Sports Med. 8, 325. Timoshenko, S. P. and Goodier, J. N. (1970) Theory ofElasticity (3rd Edn) p. 420. McGraw-Hill, New York.