Recovery of nerve · Recovery ofnerve conduction after apneumatic tourniquet cuff. Gradualrecoverywasseeninserial records, a normal response to stimulation in the thigh being obtained

Journal of Neurology, Neurosurgery, and Psychiatry, 1972, 35, 638-647

Recovery of nerve conduction after a pneumatictourniquet: observations on the hind-limb

of the baboon'T. J. FOWLER, G. DANTA, AND R. W. GILLIATT

From the Institute of Neurology, Queen Square, London

SUMMARY A small pneumatic cuff inflated around the knee was used to produce tourniquetparalysis in baboons. A cuff pressure of 1,000 mm Hg maintained for one to three hours producedparalysis of distal muscles lasting up to three months. Nerve conduction studies showed that most ofthe motor fibres to the abductor hallucis muscle were blocked at the level of the cuff and that theyconducted impulses normally in their distal parts. There was a significant correlation between theduration of compression and that of the subsequent conduction block. When tested two to threeweeks after the tourniquet, the amplitude of the response of m. abductor hallucis to nerve stimulationdistal to the cuff was usually slightly reduced compared with the precompression figure. This wasassumed to mean that a small proportion of the motor fibres had undergone Wallerian degenerationas a result of compression. Maximal motor conduction velocity was reduced in recovering nerves. Itwas also reduced when a cuff pressure of 500 mm Hg was used, which was insufficient to producepersistent conduction block. In such cases a reduced velocity without evidence of block could bedemonstrated 24 hours after compression. Ascending nerve action potentials were recorded fromthe sciatic nerve in the thigh, with stimulation at the ankle. Before compression the fastest afferentfibres had a significantly higher velocity than the fastest motor fibres in the same nerve trunk. Resultsafter compression suggested that the high-velocity afferent fibres had a susceptibility to the proceduresimilar to that of the fastest motor fibres.

In clinical practice, particularly in orthopaedicsurgery, it is well recognized that the applicationof a tourniquet to a limb is sometimes followedby paralysis which may persist for weeks ormonths. While most of the reports in the literaturedate from the period when rubber tubing orEsmarch bandages were used (Lejars, 1912;Eckhoff, 1931), occasional examples have beenreported after pneumatic tourniquets (Bruner,1951; Moldaver, 1954). In severe cases there maybe widespread Wallerian degeneration of nervesbelow the level of the tourniquet, with wasting ofmuscles, and a long delay before recovery begins.In milder cases there is a local conduction blockunder the tourniquet, without Wallerian de-generation, without wasting of muscles, andwith recovery after a few weeks. This form oflocal conduction block was investigated in cats

1 This work was supported by the Medical Research Council.

by Denny-Brown and Brenner in 1944. Theyfound that there was local demyelination of nervefibres under the tourniquet but no loss ofaxonal continuity through the lesion. Theelectrical excitability of the motor nerves distalto the tourniquet was preserved, and in mostcases power in the affected muscles recoveredwithin a few weeks. Conduction velocity duringthe recovery period was studied by Mayer andDenny-Brown (1964) who found it to be reducedat the site of the tourniquet compared with thevelocity above and below this level.

In the present experiments we have reinvesti-gated the conduction block produced by atourniquet, with special reference to the time-course of recovery. In baboons it has beenpossible to apply a pneumatic cuff rather thanthe narrower rubber tubing used by Mayer andDenny-Brown in the cat. Recovery in individual

638

Protected by copyright.

on April 19, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.35.5.638 on 1 October 1972. D

ownloaded from

Recovery of nerve conduction after a pneumatic tourniquet

animals has been followed by serial nerve con-duction studies similar to those used previouslyto study animals with experimental toxic poly-neuropathy (Hern, 1971; Hopkins and Gilliatt,1971). A brief preliminary account of ourresults has already been published (Danta,Fowler, and Gilliatt, 1971).

METHODS

All experiments were carried out on sexually maturefemale baboons (Papio papio) their body weightsranging from 8-6 to 14-9 kg. The animals were fed ona solid pellet diet (Spillers) with added fruit andmilk. Monthly injections of cyanocobalamin werealso given. As a tourniquet, a small sphygmomano-meter cuff (Accoson) designed for an infant was used.The canvas sleeve was specially reinforced to with-stand pressures of 1,000 mm Hg. The rubber bagmeasured 10 x 5 cm and was applied with its longaxis round the lower limb at the level of the knee,with the bag overlying the popliteal fossa; wheninflated its width was slightly increased to 5-5 cm.Nerve conduction was examined in the medial

popliteal branch of the sciatic nerve. For the motorstudies the stimulating cathodes were stainless steelneedles placed close to the sciatic nerve in the thigh,and close to the posterior tibial nerve below the kneeand at the ankle (Fig. 1). Muscle action potentialswere recorded from the abductor hallucis muscle,the active recording electrode being a subcutaneousneedle over the muscle belly and the remote electrodea subcutaneous needle near the base of the seconddigit.Ascending nerve action potentials were recorded

through a needle electrode close to the sciatic nervein the thigh, a remote electrode being placed over thequadriceps muscle; stimuli were delivered to theposterior tibial nerve behind the medial malleolus.Nerve and muscle action potentials were amplified bya conventional R-C coupled amplifier and displayedon one beam of a Tektronix 502 oscilloscope, theother beam being used to provide a time-scale. Thestimulus intensity was adjusted to be approximately20% greater than that necessary to produce a maxi-mal response on the oscilloscope. Single and super-imposed sweeps were photographed on 35 mm film.In addition, ascending nerve action potentials wererecorded on magnetic tape and subsequentlyaveraged in a Biomac 500 digital computer. For bothnerve and muscle action potentials, latency wasmeasured to the onset of the negative deflection.When measurements of muscle action potentialamplitude were made, the height of the negativedeflection of the action potential was used.

Anaesthesia for the application of the tourniquetand for the nerve conduction studies was providedby an initial tranquillizing dose of phencyclidine(2 mg/kg) and promazine (1 mg/kg), followed byintravenous pentobarbitone sodium (60-120 mg,depending on the duration of the experiment).To ensure that nerve cooling did not take place

during the conduction studies, animals were coveredwith a thick layer of cotton wool and the exposedlimb was warmed by a lamp. Room temperature wasmaintained above 24°C. Intramuscular temperaturein the calf was measured with a thermistor andvaried between 35.40 and 39.20 C in different experi-ments.

RESULTS

CLINICAL EFFECTS OF COMPRESSION In mostanimals a cuff pressure of 1,000 mm Hg was usedand was maintained for periods of one to threehours. In a few cases compression at 500 mm fortwo hours was tried, but these animals appearedto recover normal use of the leg within 24 hours,although electrophysiological studies indicatedthat some residual nerve damage was present. Incontrast to this, a cuff pressure of 1,000 mm Hgwas sufficient to produce paralysis of distalmuscles, which long outlasted the duration ofnerve compression.Some animals were killed for histology within

a few days or weeks, but 10 animals were keptunder observation after compression at 1,000mm Hg until substantial clinical and electro-physiological recovery had occurred, the periodof observation varying from 28 to 279 days. Intwo of the 10 cases (compression for 60 and 110minutes) the animals were only mildly disabled(B22R, B27R). Although there was foot dropand loss of toe grip, both animals started to takeweight on the affected foot within a few days,and muscle weakness disappeared within two tofour weeks. There were four severely affectedanimals (compression for 150 and 180 minutes)which showed foot drop and loss of toe gripaccompanied by a reluctance to bear weight onthe affected foot lasting about three weeks; fullmovement of the toes when gripping and climb-ing in the cage did not return for three months(Bi5, B28, B33, B34). In these animals somewasting of the muscles of the affected leg andfoot appeared, and examination under anaes-thesia at a later stage showed mild contracture,with limitation of passive plantar-flexion and

639P

rotected by copyright. on A

pril 19, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.35.5.638 on 1 O

ctober 1972. Dow

nloaded from

T. J. Fowler, G. Danta, and R. W. Gilliatt

Si S2 S3

day 1 _

dorsiflexion of the foot. The remaining animalsshowed a deficit intermediate in severity betweenthe two groups described above. In all casesappreciation of pin-prick on the affected footappeared to be preserved throughout the periodof observation.

116

lOmVr110 msec

B17

FIG. 1. Evoked muscle action potentials fromabductor hallucis muscle at different intervals after atourniquet inflated to 1,000 mm Hg round the kneefor 95 minutes. Sites of stimulating and recordingelectrodes shown below.

B2100

a

60

*L 40co/

RECOVERY OF MOTOR NERVE CONDUCTION Con-trol estimations of maximal motor conductionvelocity and response amplitude were carried outin all animals before the tourniquet was applied.No recordings were made during the period ofnerve compression or immediately after releaseof the tourniquet. Some animals were examinedafter 24 hours, but in others with severe paralysisthe first post-tourniquet recording was deferreduntil the second or third week.

All nerves compressed at 1,000 mm Hg for onehour or longer showed evidence of conductionblock. A typical result is shown in Fig. 1. It canbe seen that on the day after the tourniquet themuscle response to maximal nerve stimulation inthe thigh was reduced to a small fraction of theresponse to stimulation below the level of the

TIME IN DAYS

FIG. 2. Recovery after conduction block produced by a tourniquet in 11 nerves. Muscle actionpotentials recordedfrom abductor hallucis muscle, with motor nerve stimulation in thigh and atankle. On vertical scale, amplitude ofmuscle response to proximal stimulation as percentage ofresponse to distal stimulation. Horizontal scale, time in days after tourniquet. Animal numberand duration of tourniquet in minutes shown for each nerve.

640

49.

Protected by copyright.

on April 19, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.35.5.638 on 1 October 1972. D

ownloaded from

Recovery of nerve conduction after a pneumatic tourniquet

cuff. Gradual recovery was seen in serial records,a normal response to stimulation in the thighbeing obtained by the 116th day. Recoverycurves for 11 nerves from 10 animals are shownin Fig. 2. The number in parentheses againsteach curve is the duration of compression inminutes. The pressure in the cuff was approxi-mately 1,000 mm Hg in each case. The responseamplitude shown on the vertical scale of Fig. 2is not the absolute amplitude of the muscleresponse to nerve stimulation in the thigh, butthis figure as a percentage of the response toankle stimulation in the same experiment. Bythe use of this ratio, changes in response ampli-tude due to slight variations in the position ofthe recording electrodes on different occasionsare eliminated.

It can be seen from Fig. 2 that there was con-siderable variation in the severity of the conduc-tion block and in the rapidity of recovery indifferent animals. Some of the differencesbetween animals appeared to be related to theduration of compression. Thus the two nerveswhich were least affected were compressed foronly 60 and 110 minutes, whereas the four nerveswhich took longest to recover were subjected tocompression for 150 or 180 minutes. When thetime to 5000 recovery in Fig. 2 is matched withthe duration of the compression, using Spear-man's ranking test, this correlation is significantat the 0-01 level (r=0-82).While most of the nerve fibres subjected to

compression appeared to conduct normallydistal to the site of the cuff, there was evidencethat a small number underwent Wallerian de-generation. Thus, muscle wasting was seen inthe more severely affected animals and fibrilla-tion was present when these muscles weresampled with a coaxial needle electrode.The amount of Wallerian degeneration in each

case could be estimated roughly by comparingthe amplitude of the muscle response to nervestimulation at the ankle before the tourniquetwith the response two to four weeks after com-pression-that is, long enough after injury forWallerian degeneration to have occurred. Such acomparison is shown in Table 1.The amplitude of muscle action potentials

recorded on two successive occasions varies in arandom fashion due to slight differences in theposition of the recording electrodes, but Table 1

TABLE 1AMPLITUDE OF MUSCLE ACTION POTENTIALS RECORDEDFROM ABDUCTOR HALLUCIS MUSCLE IN RESPONSE TO MOTORNERVE STIMULATION AT ANKLE (VALUES BEFORE AND 14-24

DAYS AFTER A TOURNIQUET AT KNEE)

Nerve Duration of MAP amplitude (mV) %, changetourniquet

(min) Before After

B27R 60 20-9 24-1 + 15B27L 75 24-1 23-6 - 2B17 95 22-7 16-8 - 26B22R 110 16-3 18-6 + 14B13 120 27-8 21-4 - 23B29 150 25 5 26-2 + 3B15 150 27-7 218 - 21B16 180 27-8 16-4 - 41B28 180 22-8 20-4 - 10-5B33 180 145 136 - 6B34 180 20-9 14-5 - 30 5Mean 134-5 22-8 19 8 - 11-6

Student's t test on 11 pairs: P < 0 05.

shows a statistically significant fall in amplitudeafter nerve compression, this fall being greatestin the animals subject to the longest periods ofcompression. For these four animals (compres-sion for 180 minutes), the mean fall in responseamplitude was 22% compared with 11-6% forthe group as a whole.

In the least affected animals (compression for

a

. . . ......***....

b _

msec

FIG. 3. B26. Tourniquet at 500 mm Hg for 120minutes. Evoked muscle action potentials fromabductor hallucis muscle (a) immediately before com-pression and (b) on following day, to show increase inlatency without loss of response amplitude. Calibra-tion 10 mV. Conduction distance from stimulatingcathode in thigh to muscle was approximately 33 cmon each occasion.

641P

rotected by copyright. on A

pril 19, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.35.5.638 on 1 O

ctober 1972. Dow

nloaded from

T. J. Fowler, G. Danta, and R. W. Gilliatt

60 and 110 minutes) response amplitude wasactually slightly increased after compression,this change being within the range of variationwhich may occur in paired control observationson the same nerve (cf. Table 3 of Hopkins andGilliatt, 1971). Muscle sampling for fibrillationwas not carried out in these two animals, but theconduction studies do not suggest that the localblock at the site of the tourniquet was accom-panied by Wallerian degeneration.

In contrast with the experiments describedabove, in which a cuff pressure of 1,000 mm Hgwas used, there were three animals in which apressure of 500 mm Hg for two hours was tried.Illustrative records from one animal are shownin Fig. 3. The muscle response to nerve stimula-tion in the thigh immediately before compressionis shown above, and the response 24 hours lateris shown below. It can be seen that there waslittle change in the amplitude of the response as aresult of compression, although its latency wasincreased. Similar results were obtained for theother two nerves compressed at 500 mm Hg (seeTable 6). From this it appears that compressioninsufficient to cause a significant conductionblock may still result in a persistent conductiondelay in the affected fibres.

TABLE 2EFFECT OF TOURNIQUET AT KNEE ON MAXIMAL VELOCITY IN

MOTOR FIBRES TO ABDUCTOR HALLUCIS MUSCLE

Nerves Velocity (m/sec)(no.)

Range Mean SD

Thigh to below kneeControl nerves 26 60 4-84 5 70 3 6-7After tourniquet

early 11 13-8-44-0 29-0 10 2late 10 44 3-62 5 55 1 7 0

Below knee to ankleControl nerves 26 5383-80 0 68-7 5 3After tourniquet

early 11 56-7-79 0 65 8 8-4late 10 60-0-82-0 72-2 6-6

Table 2 shows the effect of compression onmaximal motor velocity in the 11 nerves illustra-ted in Fig. 2. In this Table the first velocityobtained after the tourniquet is labelled 'early'.In mildly affected animals this early velocity was

obtainable 24 hours after compression, whereasin severely affected cases with prolonged andcomplete conduction block, up to 45 dayselapsed before velocity could be estimated. FromTable 2 it can be seen that maximal velocitythrough the compressed segment was initiallydecreased, values for the 11 nerves ranging from13-8 to 44 0 m/sec with a mean of 29-0 m/sec.Corresponding figures for the range of controlvalues were 60-4-84-5 m/sec; mean 70 3 m/sec.The 'late' velocities shown in Table 2 were

those obtained at the last recording on eachnerve shown in Fig. 2. The time after compres-sion ranged from 28 days for the least affectednerve to 180 days for the severely affected. It canbe seen that the mean late velocity for the groupwas significantly reduced compared with thecontrol mean. Analysis of individual resultsshowed that in five nerves the velocity was stillbelow the lower limit of the control range at theend of the period of observation.

In contrast with the reduction in velocitythrough the compressed segment, the meanmaximal motor velocity in nerve distal to thecuff did not show a significant change from thecontrol mean in either the early or the late groupof observations after the tourniquet.

Figure 4 shows muscle action potentialsevoked by thigh stimulation before compression(above), and during the early recovery period

a

LJVmsec

FIG. 4. B33. Tournziquet at 1,000 mm Hg for 180minutes. Evoked muscle action potentials fromabductor hallucis muscle (a) before and (b) 73 daysafter compression, to show temporal dispersion duringearly recovery period. Calibration bar represents10 m Vfor upper and 200 ,u Vfor lower trace. Conduc-tion distance from stimulating cathode in thigh tomuscle was approximately 30 cm.

642P

rotected by copyright. on A

pril 19, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.35.5.638 on 1 O

ctober 1972. Dow

nloaded from

Recovery of nerve conzductiont after a pnzeuniatic tourniiquet

(below). The nerve had been compressed for 180minutes and the lower record was taken after aninterval of 73 days; at this time, responseamplitude was still less than 2% of the pre-compression figure. It can be seen that, in addi-tion to an increase in latency, there was temporaldispersion of the response.

Dispersion was seen only in the early stages ofrecovery of severely affected nerves. In mildlyaffected nerves and in the later stages of recoveryof severely affected nerves, the duration of themuscle action potentials did not appear to beincreased.To confirm that the reduced amplitude of the

muscle action potentials during the recoveryperiod was due to long-lasting conduction blockand not to temporal dispersion of impulses inrecovering fibres, muscle action potential areawas measured before compression and duringrecovery. In Table 3 the amplitude and the areaof the muscle response to thigh stimulation areshown as percentages of the values obtained withankle stimulation in the same experiment. Bothmeasurements refer to the negative deflection ofthe response. In control observations on 11nerves before compression, the mean amplitudeof the response to thigh stimulation was 89-10%of the amplitude obtained by ankle stimulation,whereas the corresponding figure for area was98 9,'. This suggests that in normal nerves the

TABLE 3RELATION BETWEEN AMPLITUDE AND AREA OF MUSCLEACTION POTENTIALS RECORDED BEFORE AND AFTER TOURNI-QUET (IN EACH CASE THE RESULTS OF PROXIMAL STIMULA-TION ARE GIVEN AS PERCENTAGES OF THOSE OBTAINED BY

DISTAL STIMULATION)

Nerve Before tourniquet After tourniquet

Amnp. Area Days Amtip. Area

B13 94 92 53 63 47-2B15 98 110 108 44 37 5B16 90 93-5 48 48 36-4B17 86 102-6 35 40 42-5B22R 83 104 1 49 58B27L 92-5 1015 35 59 57.5B27R 89-5 103 18 58-5 57B28 92 91 109 59 57B29 90 88-5 25 40 44-5B33 78-5 98 5 173 54 59 2B34 87 103-2 98 55 61-4

Mean 89-1 98 9 64 51-7 50-7

conduction distance between thigh and ankleallows some dispersion which results in a slightreduction in the amplitude of the response butnot in its area. To obtain corresponding valuesduring recovery, the records showing approxi-mately 5000 recovery of amplitude in the 11nerves were measured in the same way. Resultsare given in Table 3, from which it can be seenthat the mean amplitude of the muscle responseto thigh stimulation was 517% of the amplitudeobtained by ankle stimulation, whereas thecorresponding figure for area was 50-700. Thesefigures agree so closely that dispersion cannot besaid to have contributed significantly to thereduced amplitude of the muscle action poten-tials at this time. It seems that the reduction inboth amplitude and area of the thigh responsesin the recovering nerves was due to long-lastingconduction block.

ASCENDING NERVE ACTION POTENTIALS BEFOREAND AFTER COMPRESSION Ascending nerve actionpotentials (NAPs) were recorded from 21 controlnerves in 12 animals. In each case motor nervestimulation was carried out first, and maximalmotor velocity calculated. The needle electrodeplaced close to the sciatic nerve in the thigh formotor nerve stimulation was subsequently usedfor recording the ascending NAP, and thestimulating cathode at the ankle was unchangedthroughout. In this way, maximal velocity inmotor and afferent fibres could be compared inthe same nerve during the same experiment andover the same distance. Results from 21 controlnerves are shown in Table 4, from which it canbe seen that velocity in the fastest afferent fibresranged from 74-2 to 89-0 m/sec with a mean of80-9 m/sec, compared with a range of 60-0-77-0 m/sec and a mean of 68-5 m/sec for thefastest motor fibres. The difference in the meanvelocities of the two groups is significant at the0-01 level. The higher maximal velocity ofafferent fibres in this nerve is in keeping with pre-vious observations on other limb nerves in thebaboon (McLeod and Wray, 1967; Hopkins andGilliatt, 1971).Ascending NAPs were recorded from 12

nerves in 10 animals before and after compres-sion by a tourniquet. Three nerves (B30L, B33,B34) were compressed at 1,000 mm Hg for 180minutes. In the first few weeks after compression

643P

rotected by copyright. on A

pril 19, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.35.5.638 on 1 O

ctober 1972. Dow

nloaded from

T. J. Fowler, G. Danta, and R. W. Gilliatt

TABLE 4VELOCITY IN FASTEST MOTOR AND AFFERENT FIBRES OFMEDIAL POPLITEAL NERVE BETWEEN THIGH AND ANKLE

(CONTROL OBSERVATIONS ON 21 NERVES)

Velocity (mlsec)

Range Mean SD

Motor 60-0-77-0 68 5 4 0Afferent 74-2-89-0 80 9 3 5

of these three nerves, no muscle action potentials(MAPs) could be recorded from abductorhallucis in response to motor nerve stimulationin the thigh, and no NAPs could be recorded inthe thigh with stimulation at the ankle. MAPswere first recorded on the 24th, 39th, and 45thdays, and NAPs on the 38th, 60th, and 115thdays (Table 5). By the time NAPs were firstrecorded, the MAPs had recovered to 1-1, 6- 1,and 13.6% of their pre-compression amplitude.

In four nerves compressed at 1,000 mm Hg for90-120 minutes, a small MAP was present withinthe first two weeks after compression. At thistime no NAP could be recorded. Only one of theanimals was followed until the NAP reappeared.Results for this animal (B35) are shown in Table5 from which it can be seen that the NAP wasfirst recorded on the 46th day. By this time theMAP had recovered to 11% of its pre-compres-sion amplitude.

TABLE 5T'IME TO REAPPEARANCE OF MUSCLE AND NERVE ACTION

POTENTIALS AFTER TOURNIQUET

Nerve Duration of Day on which DaY on whichtourniquet (onin) MAP first seen NAPfirst seen

B30L 180 24 38B33 180 45 115B34 180 39 60B35 120 4 46

The results shown in Table 5 should not betaken to imply that afferent fibres were moreseverely affected by compression than motorfibres. It is likely that recovery of conduction inonly one or two motor fibres supplying abductor

hallucis would be sufficient to give rise to adetectable MAP. Indeed, some of the MAPsshown in Table 5 had an amplitude similar tothat of single motor unit potentials evoked bythreshold stimulation in control animals. In con-trast with this, NAPs depend upon synchronousactivity in a number of fibres, and we would notexpect to record them if only a few fibres wereconducting through the lesion. Dispersion ofimpulses in different fibres might also contributeto the late reappearance of NAPs after com-pression. Evidence that some dispersion ofascending volleys was present in the early re-covery phase of severely affected nerves can beseen in Fig. 5, which shows the post-compressionNAPs from the most severely affected nerve inour series. The rise-time of the negative deflectionof the action potential can be seen to be slightly

115 -

131-

n

0' 152-

173 -

204 -

[

[[

2msecFIG. 5. B33. Ascending nerve action potentialsrecorded from thigh, with stimulation at ankle, atdifferent intervals after a tourniquet at 1,000 mm Hgfor 180 minutes. Each tracing is the average of 128responses. Calibration 10 IL V. Conduction distance21J5-23-25 cm on different occasions. The brokenlines indicate a gradual decrease in the latency andrise-time of the potentials during recovery. (Pre-compression values for latency were 2-7 msec toinflexion, 3 3 msec to peak).

644P

rotected by copyright. on A

pril 19, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.35.5.638 on 1 O

ctober 1972. Dow

nloaded from

Recover of nierle conluctioni after a pnieumiiiatic tourn-liquiet

increased in the early stages of recovery and toreturn towards normal in later records.No comparable increase in the rise-time of

NAPs occurred in the mildly affected nerves(Fig. 6). Results from five such nerves are shown

]I

msecFIG. 6. B26. Tourniiquiet at 500 7mm Hg for 120iniitiites. Ascending nerve action potentials recordedfromtl thigh, with stimiulation at anikle (a) immlediatel/before compressioni and (b) otn following day, to showictcease in latetncyi withoult dispertsion. Calibratioln20 I V. 100 faint traces superimposed. Coniduictiondistanice 22 5-23 5 ci?,.

TABLE 6AMPLITUDE OF MUSCLE AND NERVE ACTION POTENTIALS,AND X'EIAOCITY OF FASTEST MOTOR AND AFFERENT FIBRES INFIN'E MILDLY' AFFECTED NERN'ES (IN EACH CASF AMPLITUDEAND VELOCITY IS GIVEN AS PERCENTAGE OF PRE-TOURNIQUET

FIGURE)

\e' t To ur-niq uect

Pressure Duiration

(tmittm Hg ( wtlin)

B41 L 1,000 90B32R 1,000 120B26L 500 120B30R 500 120B340L 500 120

DuYs AmoplitudC' Atax.eloc.itafWter ) (")

tourni-

quet AlAP A4P Mlotor AIthrent

4 1 1 33 5 74 0 71-66 2 5 51 0 56 5 55 01 1000 800 835 8321 103-0 81-0 85-0 67-5

915 42 5 790 78-1

in Table 6. In each case MAPs and NAPs werepresent at the first examination after compres-sion. In two nerves compressed at 1.000 mm Hg.the reduction in size of MAPs was greater thanthat of NA Ps, whereas in three nerves compressed

at 500 mm Hg the reverse was true. It is clearfrom Hopkins and Gilliatt (1971) that the ampli-tude of both MAPs and NAPs shows substantialrandom variation in successive examinations.Values for conduction velocity obtained ondifferent occasions show less variation and mighttherefore be expected to provide a more accurateindication of the severity of nerve damage. FromTable 6 it can be seen that in each case thepercentage reduction in velocity after compres-sion was similar for the fastest-conducting motorand afferent fibres. This makes it unlikely thatthere was a systematic difference in their sus-ceptibility to compression.

DISCUSSION

Our results are in good agreement with those ofDenny-Brown and Brenner (1944) who firstshowed that a tourniquet applied to the hindlimb of the cat at pressures of 450-1,200 mm Hgwould produce a persistent conduction blockwith preservation of excitability in the distalpart of the nerve. In their experiments muscleweakness usually recovered within two to threeweeks, but in a later paper by Mayer andDenny-Brown (1964) it was noted that, althoughmost animals showed marked improvement inpower by 12-16 days, reflex toe-spreading mightremain abnormal until the fourth or fifth week.In the present experiments muscle weakness andconduction block have sometimes lasted for evenlonger periods. For example, in our severelyaffected animals some weakness of toe movementlasted for approximately three months. FromFig. 2 it can be seen that in these animals few ofthe motor fibres to m. abductor hallucis hadrecovered from conduction block after twomonths and that some fibres were still blockedafter four months.

This delay in recovery is interesting in relationto clinical descriptions of tourniquet paralysis inman. Moldaver (1954) carried out serial observa-tions on two patients who developed completeparalysis of the muscles supplied by the median,ulnar, and radial nerves after surgery in whichtourniquets were used. Nerve stimulation abovethe site of the tourniquet failed to produce avisible twitch on the 38th day after injury in onepatient and on the 41 st day in the other, althoughnormal twitches were easily elicited by stimula-

645P

rotected by copyright. on A

pril 19, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.35.5.638 on 1 O

ctober 1972. Dow

nloaded from

T. J. Fowler, G. Danta, and R. W. Gilliatt

tion below this level. Recovery ofmuscle twitchesevoked by stimulation above the site of injurywas first seen after 49 days, and was complete by100 days in both cases. In some clinical accountsof patients with tourniquet paralysis, it is likelythat there was considerable Wallerian degenera-tion in addition to conduction block. In suchcases much longer delays in recovery were seen,for example in case 5 of Eckhoff (1931) and incase 2 of Richards (1954). In the three patientsdescribed by Speigel and Lewin (1945) there alsoappears to have been extensive Wallerian de-generation, and no recovery occurred during theperiod of observation.

In the present experiments the results of nervestimulation suggest that Wallerian degenerationaffected only a small proportion of the motorfibres to m. abductor hallucis, and that the slowrecovery of response amplitude which wasobserved was due to long-lasting conductionblock. It should be emphasized that regenerationafter Wallerian degeneration could not have con-tributed to the recovery of muscle action poten-tial amplitude in our study. The distance fromthe lower border of the cuff to the abductorhallucis muscle was 20-25 cm. Unpublishedexperiments on regeneration after nerve crush inthe baboon have shown that regeneration overthis distance may be expected to take threemonths; even after this time the muscle responsehas a latency greatly in excess of those observedin the present series. It is certainly remarkablethat a local block can persist for several monthsafter an episode of acute compression. Thisproblem will be discussed elsewhere in relation tothe anatomical findings (Ochoa, Fowler, andGilliatt, to be published).Although we were able to show that there was

a significant correlation between the duration ofcompression at 1000 mm Hg and the severity ofthe resulting conduction block, it is clear fromFig. 2 that individual animals varied considerablyin their susceptibility to the procedure. Forexample, compression for 180 minutes was usedin both B16 and B33; in the former, recovery ofresponse amplitude to approximately 5000 hadoccurred after 50 days, whereas in the latter ittook 167 days. Individual variation of this kindwould certainly be expected from the clinicalliterature, which contains many examples ofparalysis occurring after short periods of com-

pression which were thought to be safe on thebasis of previous experience in other patients. Inmost of these cases rubber tubing or an Esmarchbandage was used, and the pressure applied tothe limb probably varied from patient to patient(for discussion, see Hinman, 1945). Our ownresults show that when the pressure is kept con-stant, there is still considerable variation inindividual susceptibility.The effect of a pneumatic cuff on the under-

lying nerve depends not only on the cuff pressureand its duration, but also on the size of the cuff inrelation to the size of the limb, and the amount ofmuscle overlying and protecting the nerve.Denny-Brown and Brenner (1944) and Lundborg(1970) have shown that the pressure in the tissuesin the region of the nerve may be substantiallyless than that in the encircling cuff. In any casethis pressure is trivial compared with that neededto block conduction when an isolated nerve iswholly enclosed in a pressure chamber (Grund-fest, 1936). How then does pressure producelonglasting nerve damage? Does it producesecondary ischaemia by the occlusion of smallblood vessels, as suggested by Denny-Brown andBrenner (1944) and other authors? Our ownconclusion, based upon the anatomical findings,is that the major factor is not ischaemia but thepressure gradient in the nerve between its com-pressed and uncompressed parts. This results inlongitudinal movement of the axon and itsmyelin from the site of compression, and inocclusion of the nodes of Ranvier due to disloca-tion of paranodal myelin, these lesions beingconcentrated at the edges of the compressed zone(Ochoa, Danta, Fowler, and Gilliatt, 1971).The observation that conduction velocity is

reduced in nerves recovering from the effects ofcompression was first made by Mayer andDenny-Brown (1964). Their animals were studiedtwo weeks after compression, and the authorsattributed the velocity change to the demyelina-tion which was conspicuous at this time. In thepresent series, however, a reduced conductionvelocity has sometimes been recorded as early as24 hours after release of the cuff. This is tooearly for the classical appearance of demyelina-tion to be present, and some other explanationfor the velocity change must be sought. It seemslikely that both the conduction block and thereduced velocity are caused by the occlusion of

646P

rotected by copyright. on A

pril 19, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.35.5.638 on 1 O

ctober 1972. Dow

nloaded from

Recovery of nerve conduction after a pneumatic tourniquet

the nodes of Ranvier and the paranodal invagina-tion which are present during the first few daysafter compression. This problem will be discussedelsewhere in relation to the anatomical findings(Ochoa et al., to be published).

It is interesting that our results do not indicatea consistent difference in the susceptibility offast-conducting motor and afferent fibres tocompression. The relative sparing of sensationwhich occurs in many different types of humanpressure palsy may, however, be due to thesparing of small-diameter afferent fibres (Seddon,1943). While we have made no relevant physio-logical observations on this point, our anatomicalresults confirm that there is relative sparing ofsmall myelinated and unmyelinated fibres. In viewof this, it is not surprising that touch and painsensation should frequently be preserved.One general point should be made in relation

to the measurement of conduction velocitythrough a local lesion such as the one describedhere. The method which has been used formeasuring maximal motor conduction velocitymay be expected to give information about thefastest-conducting motor fibres provided thatthey can be excited by the stimulus at all levelsin the nerve trunk. If, however, the fastest fibresare blocked completely by the lesion, so thatonly slowly-conducting fibres are excited bystimuli proximal to this level, then the calculatedmaximal velocity will be misleading. Forexample, the 'early' velocities shown in Table 2were obtained at a time when most of the motorfibres to m. abductor hallucis were blocked atthe level of the lesion. Thus, although maximalshocks were used, stimulation proximal anddistal to the lesion did not excite the same fibrepopulation. For this reason the figures do notprovide valid information about velocity changesin individual fibres.

In the animals subjected to compression at500 mm Hg, however, there was little or no fallin muscle action potential amplitude and, pre-sumably, little or no conduction block. Thevelocity changes in these three nerves may thusbe accepted as a true indication of the changesoccurring in the fastest motor fibres. It can beseen from Table 6 that velocity fell to 83-5%085%0 and 7900 of the precompression value ineach case. These figures were calculated for con-duction distances of 21 5-23-5 cm between the

thigh and ankle. However, only 5-5 cm of thenerve would have been beneath the pneumaticcuff. If the velocity changes were limited to this5*5 cm length, the true velocity through thelesion might have been as low as 50-60% ofnormal, although these nerves were the leastaffected in our series. Further experiments areplanned to obtain comparable figures for singlefibres in more severely affected nerves.

We wish to thank Dr. R. G. Willison for helpfulcriticism, and Mr. P. Fitch for technical assistance.

REFERENCES

Bruner, J. M. (1951). Safety factors in the use of pneumatictourniquet for hemostasis in surgery of the hand. Journalof Bone and Joint Surgery, 33-A, 221-224.

Danta, G., Fowler, T. J., and Gilliatt, R. W. (1971). Conduc-tion block after a pneumatic tourniquet. Journal ofPhysiology, 215, 50-52P.

Denny-Brown, D., and Brenner, C. (1944). Paralysis of nerveinduced by direct pressure and by tourniquet. Archives ofNeurology and Psychiatry, 51, 1-26.

Eckhoff, N. L. (1931). Tourniquet paralysis. Lancet, 2, 343-345.

Grundfest, H. (1936). Effects of hydrostatic pressures uponthe excitability, the recovery, and the potential sequence offrog nerve. Cold Spring Harbor Symposia on QuantitativeBiology, 4, 179-187.

Hern, J. E. C. (1971). Some Effects of Experinmental Organo-phosphorus Intoxication in Primates. DM Thesis: Uni-versity of Oxford.

Hinman, F., Jr. (1945). The rational use of tourniquets.Internzational Abstracts of Surgery, 81, 357-366.

Hopkins, A. P., and Gilliatt, R. W. (1971). Motor andsensory nerve conduction velocity in the baboon: normalvalues and changes during acrylamide neuropathy. Journalof Neurology, Neurosurgery, and Psychiatry, 34, 415-426.

Lejars, F. (1912). Ce qu'il faut penser des dangers de la banded'Esmarch. Semaine Medicale, 32, 505-506.

Lundborg, G. (1970). Ischemic nerve injury. ScandinavianJournal of Plastic antd Reconstructive Suirgery, Suppl. 6.

Mayer, R. F., and Denny-Brown, D. (1964). Conductionvelocity in peripheral nerve during experimental demyelina-tion in the cat. Neutrology (Minneap.), 14, 714-726.

McLeod, J. G., and Wray, S. H. (1967). Conduction velocityand fibre diameter of the median and ulnar nerves of thebaboon. Jouirnal ofNeurology, Neuirosurgery, andPsychiatry,30, 240-247.

Moldaver, J. (1954). Tourniquet paralysis syndrome. Archivesof Suirgery, 68, 136-144.

Ochoa, J., Danta, G., Fowler, T. J., and Gilliatt, R. W. (1971).Nature of the nerve lesion caused by a pneumatic tourni-quet. Natuire, 233, 265-266.

Richards, R. L. (1954). Neurovascular lesions. In PeripheralNerve Initjries. Edited by H. J. Seddon. M.R.C. SpecialReport Series No. 282. H.M.S.O.: ILondon.

Seddon, H. J. (1943). Three types of nerve injury. Brain, 66,237-288.

Speig.l, 1. J., and Lewin, P. (1945). Tourniquet paralysis.Jouirnal of the American Medical Association, 129, 432-435.

647P

rotected by copyright. on A

pril 19, 2020 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.35.5.638 on 1 O

ctober 1972. Dow

nloaded from