spasticity - BMJ

3tournal ofNeurology, Neurosurgery, and Psychiatry 1993;56:53 1-537

The limitations of the tendon jerk as a marker ofpathological stretch reflex activity in humanspasticity

S J Fellows, H F Ross, A F Thilmann

AbstractThe motor disorders associated withhuman spasticity arise, partly from a

pathological increase in the excitabilityof muscle stretch reflexes. In clinicalpractice, reflex excitability is commonlyassessed by grading the reflex responseto a blow delivered to the tendon of a

muscle. This is a much simpler response

than the complex patterns of activitywhich may be elicited following musclestretch caused by active or passive move-ment. Changes in the biceps brachii ten-don jerk response have been followedover the first year after stroke in a groupof hemiparetic patients and comparedwith changes in short and medium laten-cy reflex responses elicited by imposedelbow flexion of initially relaxed spasticmuscle and with the development of thelate reflex responses which contribute tospastic hypertonia. A progressiveincrease in tendon jerk responses

occurred over the first year followingstroke, whereas reflex responses toimposed displacement, in particular thelate reflex responses contributing tomuscle hypertonia, reached their peakexcitability one to three months afterstroke, with a subsequent reduction inactivity. The tendon jerk reflex thereforeprovides an incomplete picture of thepathological changes in the reflexresponses in spasticity.

(7 Neurol Neurosurg Psychiatry 1993;56:531-537)

Neurologische Klinik,Alied KruppKrankenhaus, Alfried-Krupp-Strape 21,D-4300 Essen 1,GermanyS J FellowsA F Thilmann

Department ofPhysiology, UniversityofBirmingham,Birmingham B15 2TG,UKH F RossCorrespondence to:Dr FellowsReceived 13 February 1992and in final revised form27 August 1992.Accepted 11 September 1992

One of the most established tools used byneurologists in routine clinical examination is

the use of a hammer to deliver a blow to thetendon of a muscle, activating homonymousmuscle spindles which then evoke a synchro-nised reflex contraction of the muscle, thusproviding a measure of reflex excitability.'This activation was held to be mediated over

a monosynaptic pathway. Although it hasbecome apparent that oligosynaptic pathwayscontribute to this response,2 and that theafferent volley responsible is not a homoge-neous muscle spindle volley,3 the tendon jerkreflex does provide an indication of excitabili-ty in the simplest reflex pathways. A hammerblow is, however, not a stimulus which thenervous system might expect to encounter inthe course of normal behaviour. A more nat-ural stimulus is muscle stretch resulting fromimposed displacement of a limb. The

response of a muscle to such stretch, particu-larly during voluntary activation of the mus-cle, is much more complex than the simple,synchronised burst of the tendon jerk reflex.In addition to short latency components,medium and long-latency reflex componentsmay also be apparent,4 mediated over polysy-naptic spinal5 or transcortical pathways.6Furthermore, in a variety of motor disorders,for example spasticity, even relaxed muscleshows a complex pattern of response toinduced stretch.7 It is, however, more likely tobe the later, more complex responses, forexample those contributing to spastic musclehypertonia,8 which cause problems during theperformance of movements. This is oftenapparent during attempted voluntary move-ment, for example ankle dorsiflexion, whereperformance of the task may be prevented bystretch reflex activation of the antagonistsoleus muscle.9 It is therefore interesting toknow the extent to which changes in reflexexcitability, as accessed by the tendon jerkresponse, are matched by excitability changesin more functional reflexes. A study of thespastic ankle has revealed that tendon jerkresponses on the side ipsilateral to a unilateralischaemic cerebral lesion (in the area of themiddle cerebral artery) provide no indicationof a profound depression which can be seenin the stretch reflexes elicited by ankle dis-placement10 on this side, the so-called "good"side. Further, tendon jerk responses on theside contralateral to the lesion show an earlierincrease in excitability than the correspondingstretch reflexes elicited by limb displacement.

Accordingly, this study was undertaken toestablish the extent to which changes in theBiceps brachii tendon jerk response duringthe evolution of spasticity following a unilat-eral cerebral lesion reflect changes in the vari-ous components of the response of the Bicepsbrachii to imposed elbow flexion.

MethodsThe experiments were performed on twentyeight patients who had suffered a unilateralischaemic lesion in the area of the middlecerebral artery (MCA). In all cases this wasthe result of the patients' first stroke. A groupof twenty normal subjects (11 men, 9 women,aged 20-50) with no neurological abnormali-ties acted as a control. Clinical details of thepatients are given in the table. The patientswere not receiving anti-spasticity medicationduring the study, but all received regularphysiotherapy, beginning within days of their

531 on M

arch 16, 2022 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.56.5.531 on 1 May 1993. D

ownloaded from

http://jnnp.bmj.com/

Fellows, Ross, Thilmann

Table Clinical details of the hemiparetic patients

AgeCase Sex (years) Site of lesion Paresis* Tonet

< 1 month1 m 64 rMCA 2 02 m 62 rMCA 3 03 m 66 rMCA 0 04 m 59 rMCA 0 05 f 64 rMCA 3 06 f 39 IMCA 2 17 m 66 IMCA 2 18 m 40 IMCA 0 0

1-3 months9 m 52 rMCA 0 010 f 70 lMCA 2 011 f 19 IMCA 2 112 f 49 rMCA 0 313 f 75 rMCA 0 414 m 64 rMCA 1 215 m 58 IMCA 4 216 m 61 IMCA 1 317 m 78 rMCA 0 418 f 51 rMCA 4 019 m 56 IMCA 0 3

> 1 year20 m 56 IMCA 0 321 m 33 rMCA 4 222 m 47 IMCA 4 423 f 48 IMCA 4 424 f 50 IMCA 0 425 f 51 IMCA 3 426 f 40 rMCA 0 127 m 47 rMCA 4 328 m 38 rMCA 4 3

IMCA = left middle cerebral artery, rMCA = right middle cerebral artery. Localisation of thelesion was established in all cases by CT and, in most cases, by angiography and/or magnecticresonance imaging.*Paresis of the affected biceps brachii, expressed according to the MRC scale: 0 = no voluntarypower; 1 = visible contraction without affect; 2 = movement without the influence of gravity;3 = movement against gravity; 4 = movement against resistance; 5 = normal power.tTone of the spastic arm during passive extension, expressed on the Ashworth" scale:0 = normal tone; 1 = slight increase in tone, giving a "catch" when the limb is moved;2 = more marked increase in tone, but limb still easily moved; 3 = considerable increase intone-passive movement difficult; 4 = limb rigid in extension

stroke. The patients were grouped accordingto the time elapsed since stroke, the periodsstudied being less than one month, one tothree months and more than one year. Thepatients and normal subjects gave theirinformed consent to all procedures, whichhad previously been approved by the localethical committee.The subjects were seated in a stable chair,

with the arm laterally abducted and the fore-arm and hand enclosed in a plastic mould,restrained by a series of straps. The height ofthe chair was adjusted so that the shoulderwas level with, or slightly above the level ofthe arm. The mould was in turn fixed to analuminium support which was mounted on abearing adjusted to be in alignment with theaxis of the elbow joint. The mould was joinedat the wrist, via a rigid rod, to an electricmotor (Brown Boveri, MC24P, 3000W,Germany) operating under a four-quadrantcontrol system (Seidel, series 5000,Germany). This motor was capable of apply-ing elbow extension at a variety of velocities,the displacement arising from a given set ofmovement parameters being identical for allsubjects, irrespective of arm inertia or anyreflex forces developed. The parameters weresupplied by a laboratory computer (DEC,PDP1 1/73, USA). Displacement amplitudewas 300 (750 flexion to 1050), applied at 300,240, 175, 115 or 80°/s. Force was measuredby a piezo-electric transducer (Kistler 931 1A,Switzerland) placed in series with the rodconnecting the motor and the cast. Position

was monitored using a light-sensitive detector(United Detector Technology, LSC5D,USA) which followed a light source rigidlyattached to the connecting rod. Electro-myograms (EMG) were recorded from thebiceps brachii (Bb) and triceps brachii (Tb)muscles using custom-built surface electrodeswhich contained a small pre-amplifierdesigned to eliminate movement artifacts.The signals so obtained were then amplifiedand filtered (Bandwidth 2OHz-1KHz).Ten displacements were applied at each

displacement velocity. The subjects wereasked to relax their arm muscles and wereinstructed not to resist the imposed displace-ment. That they complied with these instruc-tions was established by monitoring the EMGsignals on an oscilloscope. Any trial in whichEMG was present at the onset of displace-ment was discarded. Displacement was initi-ated by the computer, which sampled, via itsADC boards, position, force and Bb and TbEMG signals with a rate of 1KHz. These sig-nals, with the motor control pulse, were alsorecorded on tape (Racal Store 7D, UK). Thecomputer rectified the EMG signals and con-structed an average from the ten displace-ments applied at each velocity. All data werewritten to disc for subsequent analysis.

For the patients, the responses were evalu-ated in terms of integrated EMG activity (inpV.s) in three segments, defined according totheir timing relative to the "normal range"(61-107ms: that is the average onset and endof the response to 3000/s displacement in nor-mal subjects: see7 8 for a full definition),namely "early" activity, appearing before thistime; "normal range" activity, that fallingwithin this interval; and "late" activity, occur-ing after this interval, up to the end of the dis-placement. For further clarity, fig 1 shows the

NormalI II SOuVN~~~~~~~5R

!f\IbJtt a(II. Spastic I25iV

100 2

25ms

Figure 1 The upper and middle traces show the rectifiedand averagedEMG responses of the relaxed Bb ofarepresentative normal subject and patient (23 in table 1),respectively, obtainedfrom ten displacements at 300°/s.The dotted lines indicate the timing of the "normal range"activity (61-107ms). The lower curve shows the positioncurve for the applied displacement.

532

on March 16, 2022 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.5.531 on 1 M

ay 1993. Dow

nloaded from


Differences in tendon jerk and stretch reflexes after stroke

Figure 2 The groupmeans of the area of theBb tendon jerk reflexEMG (in p Vs, withstandard error bar) for thenormal subjects andfor thethree groups ofpatients.

I-

0

;0:=L

.0

m

20 -

15 -

10 -

5 -

0- l

Normal < 1 month

I

1-3 months > 1 year

rectified EMG curves obtained following dis-placement at 3000/s for a representative nor-mal subject and a patient. The dotted linesindicate the timing of the "normal range". Inthe case of the "late" activity, the integral val-ues so obtained were then divided by theduration of the activity to yield a mean levelofEMG activity (in ,uV).

a)

Normals < 1 month 1-3 months > 1 year

01 IhuNormals < 1 month 1-3 months > year

Figure 3 A) The group means (with standard error bar) of the area of the "early"activity seen in the Bb EMG following an imposed extension of the initially relaxed arm of30° at 300 degls. B) The percentage ofsubjects in each group showing Bb EMG activityin this period.

Bb tendon jerk responses were obtainedwith the subject still positioned in the mould,which was fixed in the middle of the displace-ment range (that is, 900). Responses wereelicited by an experienced neurologist, usingmanual percussion of the Bb tendon.Preliminary testing had established that sucha procedure had a high degree of repro-ducibility. The blow was delivered with areflex hammer in whose head was a smallmicroswitch which closed on contact and ini-tiated computer sampling of the EMGresponses. Five responses were obtained. TheEMG record from each was rectified andthen averaged and the area of the integratedEMG record (in pV.s), its latency and dura-tion (in ms) were measured.

Statistical analysis was performed usingvariance analysis of the group data or theSpearman Rank correlation test (Statview II,Abacus Concepts, USA).

ResultsTENDON JERK RESPONSESA tendon jerk response could be evoked inall of the patients and in all but one of thenormal subjects. The group averages andstandard errors are displayed in fig 2.

In the first month after stroke, the ampli-tude of the tendon jerk response increasedabove normal values, although this increasedid not yet reach statistical significance. Thereason for this lack of significance is to befound in the wide overlap of the respectiveranges of values: the largest tendon jerkresponses obtained from the normal subjectsreached similar values to the maximumresponses of the patients tested < 1 monthafter stroke. If the percentiles of the twogroups were compared, however, it wasapparent that the patient responses were larg-er in every percentile up to the 80% level,after which the percentiles were very similar.Although the tendon jerk responses of thepatients tested < 1 month after stroke did notfall outside the range of normal values, theirresponses lay at the upper end of this range.Between 1-3 months after stroke, the tendonjerk reflex was markedly exaggerated (p <0-025). By the end of the first year afterstroke this increase had progressed still fur-ther (p < 0-001) reaching values which, onaverage, were eight times greater than thoseseen in normal subjects. At no time were sig-nificant changes observed in the latency[Mean (SD) values: normals 12(4) ms; <1month 13(4); 1-3 months 13(4); >1 year11(1)] or the duration [Mean (SD) values;normal subjects 20(6); <1 month 10(4); 1-3months 22(4)] of the tendon jerk response.

"EARLY" ACTWITYAs "early" activity (that occuring before61ms) was not observed in the normal sub-jects, its amplitude in the patients cannot becompared to normal values. Comparisonbetween the patient groups, however, can beperformed, and the resulting averages(± SEM) obtained at the fastest rate of dis-

2.0-

.0;1-m 1.6.0

a 1.2-0.8DI,

0.4-

0-

b)

100 --

be6

.0

40-

20

533



j.com/

J Neurol N


ay 1993. Dow

nloaded from



a)

2.0-

la

'w 1.6-

a 1.2-c

0.4

0.4

O -

b)

100 --

&" 80 -

.S

°A 60 -

201

Normals < 1 month 1-3 months

I

I

> I year

O I Lllfl INormals < 1 month 1-3 months > 1 year

Figure 4 A) The group means (with standard error bar) of the area of the Bb EMGactivity seen in the "normal range"following an imposed extension of the initially relaxedarm of 30° at 300 degls. B) The percentage ofsubjects in each group showing Bb EMGactivity in this period.

placement (300°/s) are shown in fig 3a.Patients showing no early activity have been,for reasons that will be made clear, excludedfrom these averages.The values in the group 1-3 months after

stroke tend to be larger than those of theother two groups, but statistically, no differ-ences were seen in the level of response ofthose patients in whom "early" activity was tobe seen. However, as is shown in fig 3b, thepercentage of subjects showing "early" activi-ty increases dramatically after the first monthof spasticity, from around 15% to 75% of thepatients, a level which thereafter increasesonly slightly. For this reason patients notshowing "early" activity were excluded fromthe group averages, as the large number ofzero values in the group < 1 month afterstroke would have given the false impressionof an increase in the level of "early" activityafter the first month of spasticity, whereas thechange lies in the number of subjectsresponding, with no significant differences inthe level of response, when present.

ACTIVITY IN THE "NORMAL RANGE"Displacement at 300°/s elicited a reflex EMGin 35% of the normal subjects. For the rea-sons given above, subjects not showing aresponse were excluded from the group aver-age. The group mean of the normals, alongwith those of the three patient groups areshown in fig 4a.No difference existed between the "normal

activity" in patients in the first month afterstroke and the amplitude of the reflex EMGof the normal subjects. After the first month,the level of response increased, but this didnot reach significant levels. If the percentageof subjects showing a response is calculated,however, it is apparent (fig 4b) that anincrease occurs within weeks of a stroke, witha subsequent progression until, by the end ofthe first year after stroke, all patients demon-strated activity in the "normal range". Thusthe alteration in this response in spastic sub-jects consists not of a significant increase inthe level of response, but rather in the abnor-mally high occurrence of this response, whichbecomes ever more marked over the course ofthe first year after stroke.

Associated with this increase in the numberof patients showing a response is a steadydecrease in the lowest displacement velocityat which a response is first evident.

Figure 5 shows the distributions of thereflex threshold of activity in the "normalrange" for the normal subjects and for thethree patient groups. It may be seen thatwhile no normal subject showed a reflexresponse with displacement velocities lowerthan 175°/s, by the end of the first year afterstroke a large percentage (78%) responded todisplacement at 115°/s, and a minority stillshowed activity with stretch velocities as lowas 80°/s.

"LATE" ACTIVITYAs would be expected from a previous study,8all patients in the groups tested 1-3 monthsor >1 year after stroke who were clinicallyassessed to have raised muscle tone in the Bbresponded to imposed stretch with "late"reflex EMG activity. This began immediatelyafter the "normal range" (107ms) and termi-nated at or around the end of the displace-ment. The mean level of this EMG activityshowed a highly significant positive linearcorrelation with displacement velocity. As haspreviously been shown,8 the slope of theregression plot between mean Bb EMG andvelocity represents the gain of the reflexmechanism underlying this activity.Additionally, the mean level of the "late"EMG activity showed a significant linear pos-itive correlation to the passive muscle hyper-tonia clinically assessed in these patients onthe Ashworth Scale" (1-3 months r = 088,p = 0O008; >1 year r = 07, p = 0048). In thepatient group tested in the first month afterstroke, although hypertonia was positivelyscored on the Ashworth scale in only twosubjects, several others also showed some"late" Bb EMG activity. In these cases, how-ever, in common with the two subjects in

534 on M


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



14 -

1 10-

E

8-

4-

2-

0

Normals

<1 month

1-3 months

> Iyear

.I[ F~~~~~~~~~~~~~~~~80 115 175 240 300 > 300

Reflex threshold (deg/s)

Figure S The reflex thresholds of the normal subjects and the three patient groups.Threshold is defined as the lowest of the five displacement velocities applied which elicited aresponse in a given subject. >300 indicates that no reflex response was obtained with thehighest velocity used.

whom hypertonia was scored clinically (seetable), this activity was at a very low level,terminated well before the end of the dis-placement and, in contrast to the "late"responses of the patients in the other twogroups, did not appear with stretch velocitiesbelow 175°/s. It is thus questionable if the"late" activity seen in these patients testedless than one month after stroke is strictlycomparable with the activity seen in the otherpatient groups. For the sake of completeness,however, all subjects in the earliest patient

Figure 6 The groupmeans (with standard errorbar) of the average level ofBb EMG activity duringthe 'late" response evokedfollowing an imposedextension of the initiallyrelaxed arm of30' at 300degls.

357

0 30-

25

20 -

1S-

10-

5

0

< 1 month

I

1-3 monoiths > 1 year

group showing late Bb EMG (n = 5) wereincluded in the statistical analysis.

Figure 6 shows the mean level of Bb EMG(± SEM) during the "late" response elicitedduring displacement at 300°/s for the threegroups. It is apparent that in the first monthof spasticity, only a very low level of Bb EMGactivity was evoke-1. Between the first andthird month after stroke, however, the level of"late" Bb EMG increased considerably(p < 0-01). Such an increase is in line withthe behaviour of the tendon jerk reflex at thistime, but whereas the tendon jerk subse-quently showed a further increase in patientsa year or more after stroke, it may be seen infig 6 that the "late" Bb EMG obtained inthese patients was considerably reduced fromthe level of activity seen in the patients 1-3months after stroke (p < 0 01). Thus a dis-parity exists between the behaviour of thetendon jerk reflex and the late reflex EMGactivity which contributes to one of the majormotor disturbances associated with spasticity,namely muscle hypertonia.

INTRA-INDIVIDUAL DIFFERENCESThe previous sections have established that,for group mean values, clear differences existin the behaviour of tendon jerk reflexes and avariety of reflexes elicited by limb displace-ment. It is valid to ask therefore if the tendonjerk of individual patients reflected the levelof response of the reflex responses to imposeddisplacement. To answer this question acomparison was made between the size of thedifferent reflex components in the variousgroups. It was found that for the patients test-ed less than one month or more than one yearafter stroke, the magnitude of the tendon jerkresponse showed no significant correlationwith the magnitude of any of the reflex com-ponents obtained by limb displacement (no pvalues smaller than 0 383). For the grouptested between 1-3 months after stroke, themagnitude of the tendon jerk response waspositively correlated with the "early" activity(p = 0-016) and the activity elicited in the"normal range" (p = 0001). No significantcorrelation, however, was apparent with themean level of the "late" reflex activity (p =0127).

DiscussionThe results of this study agree with an earlierfinding obtained at the hemiparetic ankle,10that the changes in the tendon jerk responseduring the development of spasticity do notinevitably reflect alterations in more naturallyelicited stretch reflex responses, such as thoseresulting from limb displacement. This find-ing has important implications not only forclinical assessment of spasticity, but also forstudies of the pathophysiological mechanismsunderlying spasticity and for studies intendedto evaluate the efficacy of therapeutic inter-vention, whether pharmacological or physical.The finding that the tendon jerk response

of the patient group tested <1 month afterstroke was not significantly elevated over nor-

535



j.com/

J Neurol N


ay 1993. Dow

nloaded from



mal values might, at first sight, be consideredsurprising from a clinical viewpoint: it is gen-erally accepted that exaggerated tendon jerksare first apparent within days, rather thanweeks of the causative lesion. Furthermore,experimental studies in the rat and cat havedemonstrated that monosynaptic reflexes aresignificantly increased almost immediatelyfollowing a spinal cord lesion.1213 It must beremembered, however, that clinical assess-ment of elevated tendon jerk reflexes is per-formed as a comparison of the two sides of apatient. A recent study has emphasised, how-ever, that this cannot be considered to be thesame as a comparison against normal values,on the grounds that the Bb tendon jerkresponse ipsilateral to a strictly unilateralcerebral lesion is significantly reduced fromnormal values.7 Thus the comparison of theBb tendon jerk responses on both sides of apatient with a unilateral cerebral lesion maylead to an exaggerated impression of theincrease on the side contralateral to thelesion, relative to the actual increase over nor-mal control values. In addition, serial studiesof H-reflex changes following spinal cordinjury in humans have revealed that signifi-cant increase in a measure of H-reflexexcitability, the H/M ratio, are first apparent4-6 months after injury.'4 The same grouphave also shown in experiments on the cat'5that the increase in monosynaptic reflexexcitability following spinal cord transectionis biphasic, with an early increase 1-4 daysafter transection, followed by a more signifi-cant rise between 2 and 4 weeks later.Although the situation following spinal tran-section may not be fully comparable with thatpertaining after stroke, this biphasic patternof increase may be reflected in the results ofthe present study. The earliest increase maybe reflected in the observed tendency for theBb tendon jerk response amplitudes ofpatients tested < 1 month after stroke to clus-ter at the upper end of the normal range. Thesecond, more marked phase of increase maythen be apparent as the significant increase ofthe tendon jerk response seen in the patientstested 1-3 months after stroke.The progressive increase in the magnitude

of the tendon jerk response over the course ofthe first year after stroke does not reflect themore complicated changes in the "late" EMGresponses contributing to spastic musclehypertonia, which decreases after reaching amaximum 1-3 months after stroke, nor therelatively unaltered state of shorter latencyreflexes elicited by limb displacement. Thesteady increase is, however, mirrored by thenumber of patients in whom shorter latencyresponses to limb displacement were to beseen. At first sight the connection between anincrease in tendon jerk response magnitudeand the frequency of occurrence of responseswhich do not show parallel changes in magni-tude may not seem obvious. If, however, thelikely pathways of the various responses areconsidered, a possible explanation becomesapparent. While the tendon jerk response ismediated over monosynaptic and oligosynap-

tic connections,2 the latency oi the reflexactivity observed in the "normal range"would indicate that there exists a significantinterneuronal contribution to this response.Although, given its short onset latency (asearly as 25ms after the onset of displace-ment), the "early" activity is likely to bemediated at least in part over monosynapticand oligosynaptic connections, the absence ofa progressive increase in the magnitude ofthis activity would indicate that other connec-tions, at a more complex level than that sub-serving the tendon jerk response and subjectto a wider range of influences, are alsoinvolved in this response. It is possible there-fore that the increase in the tendon jerkresponse reflects excitability changes at a verybasic level of the motor system. Changes inreceptor excitability, pre-synaptic inhibitionof afferent transmission and motor neuronexcitability have all been considered as possi-ble mechanisms of spastic hyperreflexia.Clear evidence has, however, only been foundfor the second of these suggestions: the inhi-bition of H-reflex responses by tendon vibra-tion, which is thought to be mediated by apre-synaptic mechanism, has been shown tobe greatly reduced in spastic patients.'6 17Microneurographic recordings of musclespindle afferents in spastic humans havefailed, as yet, to demonstrate any evidence ofreceptor hyperexcitability in the relaxedstate.'8 Such a receptor hyperexcitability, forexample in the muscle spindle endings would,however, be compatible with the observedincreases in the amplitude of monosynapticreflex responses and could also explain thereduction in threshold and increase in the latereflex components mediated over pathwayswith greater interneuronal involvement.Under this schema, an increase in receptorsensitivity would lead to the drop in thethreshold of these later components and to anincrease in the number of patients in whomthey were observed. The different timecourseof alterations in the magnitude of the latercomponents would then reflect additionalinfluences operating on the more complexparts of their pathways (that is, the interneu-rons), which are not apparent at a mono- oroligo-synaptic level. Until evidence forincreased receptor sensitivity is provided,however, such an explanation must remainspeculative.An obvious question which arises from this

disparity between the behaviour of the tendonjerk reflex and more naturally elicited reflexresponses is, what relevance does it have forthe patient and the clinician? Clearly, in clini-cal practice a diagnosis of spasticity is unlikelyto be made on the basis of exaggerated ten-don jerk reflexes alone, but in many studiesattempting to evaluate the efficacy of physicalor pharmaceutical intervention, evaluation ofchanges in reflex behaviour are often made onjust such a basis, from either tendon jerk orH-reflex studies. The results of our studyindicate that investigations restricting them-selves to these simple peripheral pathwayswill either fail to reveal, or run the risk of mis-

536 on M


http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



interpreting, the complex changes occurringin more physiological reflex responses, whichcause the patient far more problems thanalterations in a non-physiological parametersuch as the tendon jerk reflex.9 Similarly,although it has been shown that short-latency("monosynaptic") reflexes generated in spas-tic patients during gait contribute little or noresisting force to an imposed displacement,'9it would be wrong, on the basis of the behav-iour of monosynaptic reflex responses alone,to state that hyperreflexia plays no part in themotor disability of the spastic subject.20 It isapparent that future studies must be concen-trated on the more complex, movementevoked reflex responses of the motor system ifa correct picture of the pathophysiology ofspasticity is to be successfully achieved, and iflikely modes of therapy are to be assessed andthe success of such intervention correctlyevaluated.

This work was supported by the DeutscheForschungsgemeinschaft. We thank Elisabeth Garms for herexperimental assistance for part of this study and Professor JNoth and Drs Domges, Schwarz and T6pper for their helpfulcomments on the manuscript.

1 Lansaka DJ. The history of reflex hammers. Neurology1989;39: 1542-9.

2 Burke D, Gandevia SC, McKeon B. Monosynaptic andoligosynaptic contributions to human ankle jerk and H-reflex. JNeurophysiol 1984;52:435-48.

3 Burke D, Gandevia SC, McKeon B. The afferent volleysresponsible for spinal proprioceptive reflexes in man. JPhysiol 1983;339:535-52.

4 Lee RG, Tatton WG. Motor responses to sudden limbdisplacements in primates with specific CNS lesions andin human patients with motor system disorders. Can JNeurol Sci 1975;2:285-93.

5 Hultbom H, Wigstrom H. Motor response with longlatency and maintained duration evoked by activity in Iaafferents. In: Desmedt JE, ed. Super and supraspinalmechanisms of voluntary motor control and locomotion.Progress in dinical Neurophysiology 8. Basel: Karger,1980:99-116.

6 Marsden CD, Merton PA, Morton HB. Stretch reflex andservo action in a variety of human muscles. J Physiol1976;259:531-60.

7 Thilmann AF, Fellows SJ, Garms E. Pathological stretchreflexes on the "good" side of hemiparetic patients. JNeurol Neurosurg Psychiatry 1990;53:208-14.

8 Thilmann AF, Fellows SJ, Garms E. The mechanism ofspastic muscle hypertonus: variation in reflex gain overthe time course of spasticity. Brain 1991;114:233-44.

9 Corcos DM, Gottlieb GL, Penn RD, Myklebust B,Agarwal GC. Movement deficits caused by hyperex-citable stretch reflexes in spastic humans. Brain1986;109: 1043-58.

10 Thilmann AF, Fellows SJ. The time-course of bilateralchanges in the reflex excitability of relaxed triceps suraemuscle in human hemiparetic spasticity. J Neurol199 1;238:293-8.

11 Ashworth B. Preliminary trial of Carisoprodol in multiplesclerosis. Practitioner 1964;192:540-2.

12 Maimsten J. Time course of segmental reflex changes afterchronic spinal cord hemisection in the rat. Acta PhysiolScand 1983;119:435-43.

13 Chambers WW, Lui C-N, McCouch GP, D'Aquili E.Descending tracts and spinal shock in the cat. Brain1966;89:377-90.

14 Little LW, Harlar EM. H-reflex changes following spinalcord injury. Arch Phys Med Rehabil 1985;66:19-22.

15 Little LW. Serial recording of reflexes after feline spinalcord transection. Exp Neurol 1986;93:510-21.

16 Ashby P, Verrier M, Carleton S, Somerville J. Vibratoryinhibition of the monosynaptic reflex and presynapticinhibition in man. In: Feldman RG, Young RR, KoellaWP, eds. Spasticity: disordered motor control. Chicago:Year Book Medical, 1980:335-44.

17 fles JF, Roberts RC. Presynaptic inhibition of monosynap-tic reflexes in the lower limbs of subjects with uppermotoneuron disease J Neurol Neurosurg Psychiatry1986;49:937-44.

18 Hagbarth K-E, Wallin G, Lofstedt L. Muscle spindleresponses to stretch in normal and spastic subjects.ScandJ rehab Med 1979;5:156-9.

19 Berger W, Horstmann GA, Dietz V. Tension developmentand muscle activation in the leg during gait in spastichemiparesis: the independence of muscle hypertoniaand exaggerated stretch reflexes. J Neurol NeurosurgPsychiatry 1984;47:1029-33.

20 Dietz V. Spastik: Therapie der gesteigerten Reflexe oderder Bewegungsstorung? Nervenartz 1990;61:581-6.

537



j.com/

J Neurol N


ay 1993. Dow

nloaded from


Documents

spasticity - BMJ