Joiurnial of Neurology, Neurosurgery, anid PsYchiatrY 1981 ;44 :759-767 The pathogenesis of pneumatic tourniquet paralysis in man STEPHEN K YATES, LAWRENCE N HURST, WILLIAM F BROWN From the Department of Clinical Neurological Sciences, University Hospital, London, Ontario, Canada SUMMARY The relative importance of ischaemic and direct mechanical injury to nerves compressed by a tourniquet, in the pathogenesis of tourniquet paralysis in man has not been established. To investigate this question, conduction in ulnar or median nerve fibres has been measured in healthy subjects both at the level of the pneumatic tourniquet and distal to the tourniquet. Measurement was prior to, for the period of tourniquet inflation, and following release of the tourniquet. The earliest conduction delays and block were observed at the level of the tourniquet, particularly across the proximal tourniquet border zone. However, a proximal to distal progression in conduction abnor- malities distal to the tourniquet suggested that the earlier conduction abnormalities at the level of the tourniquet were primarily ischaemic in origin. Mechanical compression, however, probably contributed to disproportionate conduction delays and blocks across the border zones of the tourniquet. The pneumatic tourniquet is commonly employed by surgeons to provide a bloodless surgical field in the limbs. The complication rate at the University Hospital in London, Ontario has been estimated to be 1 in 5,000 or less. Rarely, paralysis, loss of sensation, or autonomic abnormalities may persist for periods of several months, even when the tourniquet is employed at recommended pressure levels and inflation times.'-3 Local conduction delays have been observed in peripheral nerves at the level of the tourniquet border zone in patients with post-tourniquet paralysis.3 These local conduction abnormalities may be the functional equivalent of the structural abnormalities observed at the proximal-distal border zones in experimental tourniquet lesions.4 These occurred at inflation times and pressures clearly beyond those employed in man. It is not known, however, what the relative importance of ischaemia and com- pression are to the abnormalities in nerve function at the level of the tourniquet, particularly at pressure levels and inflation times normally employed by surgeons. Experiments were therefore devised in an attempt to quantitate the relative importance of ischaemia and mechanical compression to the Address for reprint requests: Dr WF Brown, University Hospital, 339 Windermere Rd, London, Ontario, Canada, N6A 5A5. Accepted 8 June 1981 abnormalities in nerve function during and in the early period following tourniquet inflation in man. Methods These investigations were carried out during mammo- plastic reductions. The protocol was reviewed and ap- proved by the human experimentation committee at the University of Western Ontario. Informed consent was obtained from the patients. The operations were carried out under general anaesthesia but without the use of neuromuscular blocking agents. The Kidde Model 400 pneumatic tourniquet was employed (80 mm wide). This tourniquet employs a plastic cuff insert external to the inflatable cuff. The maximum inflation pressure employed was 300 mm Hg. Tourniquet inflation times varied between A-1 hour. Special care was taken to avoid wrink- ling the skin beneath the tourniquet. For most of the experiments only one tourniquet was employed, centred at the level of the mid-upper arm, but in later experiments two tourniquets were employed, one at the level of the mid-forearm and the other about the upper arm. Electi/ophtysiological methods The majority of the experiments were carried out on the ulnar nerve-hypothenar muscle complex (fig 1). The ulnar nerve was stimulated at the level of the wrist and proximal and distal to the upper arm tourniquet in the earliest experiments. Later, extra stimulating electrodes under the tourniquet made it possible to measure the conduc- tion times across the sub-tourniquet and proximal and distal tourniquet border zones (fig 3). For these later investigations, a special multi-electrode was constructed 759
Transcript
Joiurnial of Neurology, Neurosurgery, anid PsYchiatrY 1981 ;44 :759-767
The pathogenesis of pneumatic tourniquet paralysisin man
STEPHEN K YATES, LAWRENCE N HURST, WILLIAM F BROWN
From the Department of Clinical Neurological Sciences, University Hospital, London, Ontario, Canada
SUMMARY The relative importance of ischaemic and direct mechanical injury to nerves compressedby a tourniquet, in the pathogenesis of tourniquet paralysis in man has not been established. Toinvestigate this question, conduction in ulnar or median nerve fibres has been measured in healthysubjects both at the level of the pneumatic tourniquet and distal to the tourniquet. Measurement was
prior to, for the period of tourniquet inflation, and following release of the tourniquet. The earliestconduction delays and block were observed at the level of the tourniquet, particularly across theproximal tourniquet border zone. However, a proximal to distal progression in conduction abnor-malities distal to the tourniquet suggested that the earlier conduction abnormalities at the level ofthe tourniquet were primarily ischaemic in origin. Mechanical compression, however, probablycontributed to disproportionate conduction delays and blocks across the border zones of thetourniquet.
The pneumatic tourniquet is commonly employedby surgeons to provide a bloodless surgical field inthe limbs. The complication rate at the UniversityHospital in London, Ontario has been estimated tobe 1 in 5,000 or less. Rarely, paralysis, loss ofsensation, or autonomic abnormalities may persistfor periods of several months, even when thetourniquet is employed at recommended pressurelevels and inflation times.'-3
Local conduction delays have been observed inperipheral nerves at the level of the tourniquet borderzone in patients with post-tourniquet paralysis.3These local conduction abnormalities may be thefunctional equivalent of the structural abnormalitiesobserved at the proximal-distal border zones inexperimental tourniquet lesions.4 These occurred atinflation times and pressures clearly beyond thoseemployed in man. It is not known, however, whatthe relative importance of ischaemia and com-pression are to the abnormalities in nerve function atthe level of the tourniquet, particularly at pressurelevels and inflation times normally employed bysurgeons. Experiments were therefore devised in anattempt to quantitate the relative importance ofischaemia and mechanical compression to the
Address for reprint requests: Dr WF Brown, UniversityHospital, 339 Windermere Rd, London, Ontario, Canada,N6A 5A5.
Accepted 8 June 1981
abnormalities in nerve function during and in theearly period following tourniquet inflation in man.
Methods
These investigations were carried out during mammo-plastic reductions. The protocol was reviewed and ap-proved by the human experimentation committee at theUniversity of Western Ontario. Informed consent wasobtained from the patients. The operations were carriedout under general anaesthesia but without the use ofneuromuscular blocking agents. The Kidde Model 400pneumatic tourniquet was employed (80 mm wide). Thistourniquet employs a plastic cuff insert external to theinflatable cuff. The maximum inflation pressure employedwas 300 mm Hg. Tourniquet inflation times variedbetween A-1 hour. Special care was taken to avoid wrink-ling the skin beneath the tourniquet. For most of theexperiments only one tourniquet was employed, centredat the level of the mid-upper arm, but in later experimentstwo tourniquets were employed, one at the level of themid-forearm and the other about the upper arm.
Electi/ophtysiological methodsThe majority of the experiments were carried out on theulnar nerve-hypothenar muscle complex (fig 1). The ulnarnerve was stimulated at the level of the wrist and proximaland distal to the upper arm tourniquet in the earliestexperiments. Later, extra stimulating electrodes underthe tourniquet made it possible to measure the conduc-tion times across the sub-tourniquet and proximal anddistal tourniquet border zones (fig 3). For these laterinvestigations, a special multi-electrode was constructed
759
760Yates, Hurst, Brown
orsimplified *
Rm<
Rs <, R:K Rs
T
Fig I Left: The general arrangement of the electrodesand the position oJ the pneumatic tourniquet about theupper arm. Right: Corresponding schematic outline of therelative electrode positions and the tourniquet.S = Stimulating electrode (cathode) positions, Rm =Recording electrodes (surface) over hypothenar musclegroup, R, = Recording electrodes (DISA 13L69)positions about the Vth digit (G, proximal), T =Tourniquet, N. = Nerve-ulnar, HT = Hypothenar.
which incorporated 4 x 10 mm silver strip electrodes onthe inner surface. The interelectrode distance across thetourniquet border zones was 50 mm and between thesub-tourniquet electrodes, 30 mm. A large silver plateover the upper back was employed as the anode. In theearly experiments needle electrodes were used to stimu-late the ulnar or median nerves proximal and distal tothe tourniquet but these were replaced by surface elec-trodes in the majority of the subsequent investigations.In all these experiments special care was taken to ensurethat the brief (0-1 ms or less) isolated stimulus pulseswere just supramaximal and that the stimulating elec-trodes were placed directly over the respective test nerve.The latter was achieved by placing the electrodes wherethe maximum response was obtained at the least stimulusintensity. These criteria were rechecked after the tourni-quet was placed about the forearm and in the earlyperiod following tourniquet inflation to minimise shiftsin the location of the stimulating electrodes in relationto the nerve.The temperature of the limb distal to the tourniquet
was monitored by an intra-muscular thermistor electrode.Once all electrodes were in place, the limb was insulatedby cotton wool, etc, to reduce the heat loss and therebyminimise the conduction changes which could resultfrom cooling of the nerve and muscles. Testing wascarried out prior to tourniquet inflation, at five minute
intervals or less following tourniquet inflation andcontinued for up to one hour after tourniquet release.The least conduction times (or conversely the maximumconduction velocities) and the degree of conduction blockwere measured across the level of the tourniquet and
-S distal to the tourniquet. the latter in the pure ischaemiczone. The "M" response was recorded by surfaceelectrodes. The position of the stigmatic electrode wasadjusted to obtain the maximum peak-to-peak voltage(p-pV), least rise time and initial negative deflection for
-S the thenar (T) or hypothenar (HT) muscle groupsinnervated by the respective ulnar or median nerves. Thedegree of conduction block was estimated by changes inthe maximum "M" response p-pV or area. Area was
Nu measured only in the earliest experiments because it waslearned that there was little change in p-pV duration inthe "M" response, that is, little evidence of temporal
c: dispersion in the "M" response (figs 2, 3).
T
Tourniquettimne (mins) /'7
Pre test
On 0 0------------------
S3-5S2
- Si
HT)
26
J10mV28 -
34
38
14460
64
0 5 10 15 20 25Time (ms)
Fig 2 Left: Schematic of the relative positions of thestimulating electrodes at the levels of the wrist (Sj) justdistal to the tourniquet (S2) and proximal to the tourni-quet (S3) about the mid-upper arm. Right: Selectedrecords to illustrate the changes in the hypothenar(HT) "M" response to stimulation at the S1, S2 and S3levels at various times prior to and after tourniquetinflation andfollowing release of the tourniquet. Note(1) the earliest conduction block across the tourniquetsegment (between S2 and S3), (2) persistence of the "M"response evoked by stimulation at S1, (3) the overallcascade beginning at the most proximal level ofstimulation and moving toward the most distalstimulation level and (4) the relative degree of conductionblock and delay.
k
760
HT Rm
The pathogentesis ofpnieumatic tourniiquet paralysis inman7t
Pre test
S5On * 0
-4 15S3
S2 20
32
60Off *
]lOmV
25
HT)
1 5
-l -L0 5 10 15 20 25
Time ( ms)
Fig 3 Left: Schematic of the elative positions of thestimulating electrodes including S at the level of thewrist, S2 just distal to the tourniquet, S3 and S4 beneaththe tourniquet and 55 just proximal to the tourniquet.Tourniquet about the mid-upper arm. Right: The changesin the hypothenar (HT) "M" response at various timesafter tourniquet inflation and release in relation to thepre-test observations. Note (1) the earliest conductionblock across the proximal tourniquet border zone, (2)the early disproportionate conduction delay across thedistal tourniquet border zone in relation to the sub-tourniquet segment, (3) the overall proximal to distalcascade in the degree of conduction block and (4)following tourniquet release, the disproportionateconduction delay across the distal tourniquet borderzone at 2 5 minutes and persistent conduction blockacross the segment even at 15 minutes.
Results
(1) THE RELATIVE CHANGES IN
CONDUCTION ACROSS THE LEVEL OF THETOURNIQUET AND DISTAL TO THE
TOURNIQUETThe earliest conduction block was observed acrossthe level of the tourniquet at 10-20 minutes in HTmotor fibres (fig 2). Only later, between 20-30minutes, was conduction block evident distal to thetourniquet and only by 40 minutes over the mostdistal segment, beyond the level of the wrist. The
largest relative increase in conduction time wasobserved across the tourniquet zone. Equivalentreductions in the maximum conduction velocitywere observed only later, distal to the tourniquet;corresponding increases in the motor terminallatency being delayed the longest (figs 2, 3). Justprior to complete impulse block, the maximumconduction velocity (MCV) was less than one halfthe pre-tourniquet values. These reductions werenot accounted for by lower temperatures in the armbecause only a 1-3 degree centigrade loss wasobserved provided the limb was properly insulated.
(2) CHANGES IN CONDUCTION ACROSS THESUB-TOURNIQUET SEGMENT RELATIVE TOTHE PROXIMAL AND DISTAL TOURNIQUETBORDER ZONES
(i) Motor fibresIn HT motor fibres the earliest (less than 15 minutes)and largest increases in conduction time wereobserved across the proximal tourniquet borderzone (figs 3 and 4A-B). In some subjects (fig 3)relatively larger increases in the conduction timeacross the distal tourniquet border zone in relationto the sub-tourniquet segment were observed. Thiswas not observed in all subjects (fig 4). Conductionblock was evident earlier (15-20 minutes) and wasfirst complete across the proximal tourniquet borderzone. The observation that little or no change in p-pduration accompanied the reduction in the p-pVmeant that it was unlikely the latter was a product,in part, of temporal dispersion of the "M" responseor preferential block of the fastest or slowest con-duction velocity fibres.
(ii) Sensory fibresThe earliest and maximum conduction blocks wereobserved across the tourniquet border zones, particu-larly the proximal border zone. This was evidentwithin less than 10 minutes of tourniquet inflation.Because of the shorter duration of single nervefibre action potentials (SNAP) (< 5 ms) comparedto motor unit potentials (10-20 ms) and the lowerrelative voltage (< 0-1 mV) of the maximum SNAPrelative to even a single motor unit potential(< 0 5 mV) it was much harder to quantitate thecontribution of temporal dispersion to reductions inthe SNAP p-pV. However, whatever the relativecontribution of conduction block and temporaldispersion to the observed changes, a disproportion-ate reduction in the SNAP p-pV across the tourni-quet border zones in relation to the sub-tourniquetsegment was observed in the first 20 minutes.
T I
I
761
'..u,V,/I
A- /- , i.
ll-.
---------------------------------------
--L.-
Yates, Hurst, Brown
Tourniquet0 g4o0 - o,
0
'o\
*~~0
---o F
o15020
026
\ 28
32
Do
80 0 32
50o28
40 :o00-/ o2
o 0 ~~~~~~~~~-o2020 0 0== 015
00 _ 0----- 0 -o----o--- 0 p
80 I I.O IIo
c 0o03002
E 0.025-
> 0 020E(0 Ea a 0015o .c K
3 0-005(0
0
Di00
3.Ov
2
0
100 200 300 400 500
Distance from motor point (mm )
Fig 4 Electrode locations relative to an upper arm
tourniquet idential to fig 3. Upper: Plot of the percentagechanges at various times (in minutes) in the maximumhypothenar "M" response (compared to the initial values)evoked by stimulation of the ulnar nerve proximal,beneath and distal to the tourniquet. Lower: Correspond-ing percentage changes in least latencies of thehypothenar "M" response evoked by stimulation at thevarious levels compared to their respective initial values.The parallel vertical lines illustrate the positions of theproximal and distal tourniquet border zones in relationto the electrode positions. The distance from thehypothenar motor point to the stimulating electrode(in mm) is plotted on the X axis. Note (1) the earlydisproportionate conduction delays across the tourniquetborder zones and (2) the earliest conduction block across
the proximal, and in this patient, sub-tourniquet zones.
(3) PROXIMAL TO DISTAL PROGRESSION
IN THE CONDUCTION ABNORMALITIESThe above observations established that the earliestand maximum functional abnormalities were obser-ved at the level of the tourniquet; specifically across
the proximal tourniquet border zone. These observa-tions were initially interpreted to indicate a primarymechanical rather than ischaemic factor in thepathogenesis of the functional abnormalities at thelevel of the tourniquet. It was obvious, however,
10C/, 15 20/
480 440 400 360 320 240 160 80
istonce (mm) to recording elecirode Vth digit
Fig 5 Schematic as per fig 1. Plot of the negative peakvoltage (- p V) in mV of the antidromic fifth digitsensory nerve action potential (AD Vth Digit SNAP) at
various times expressed in minutes following inflationof the pneumatic tourniquet about the upper arm.
Ulnar nerve stimulation was carried out at four levelsin the upper arm to measure the changes in the SNAP- p V across the proximal and distal tourniquet borderzones and across the sub-tourniquet segment. The ulnarnerve was also stimulated at the level of the wrist. Theposition of the tourniquet border zone relative to thestimulating electrodes is shown by the interruptedvertical lines. Note (1) the different voltage scales forthe SNAP values in response to stimulation at the wristand the other stimulation points at the upper arm, (2) thedistance between the recording electrode (G1) and thestimulating cathode is shown on the X axis in mm. Thescale factor is changed at 320 mm. The earliest reductionin SNAP voltage was observed at 5 min across theproximal tourniquet border zone and disproportionatereductions in SNAP voltage were obvious across bothtourniquet zones by 10 min. The interrupted extensionsrepresent what the expected changes with distance mighthave been without conduction block or temporal disper-sion. By 15 minutes there was complete block across thesub-tourniquet and proximal border zone segments.
that there was a proximal to distal progression in theincreases in conduction time and conduction blockin the ischaemic zone distal to the tourniquet in bothmotor and sensory fibres. For example, the conduc-tion time increases and magnitude of the conductionblock at equivalent ischaemic times were more
abnormal across the forearm than in the terminalwrist to muscle segment (figs 1, 2, 3, 4). To illus-trate: in fig 2, at 28 minutes the increase in conductiontime across the forearm segment was 22% but therewas only a 13% increase in the motor terminallatency, the corresponding increase in conductiontime across the tourniquet segment being 53 %.There was, therefore, evidence for a proximal to
distal progression in the degree of conduction blockand conduction delay both in motor and sensory
fibres. This in turn implied that the degree of con-
100a)
0> 800
0CL 608
x 40
c 20E
.050 0
0 040 Lto
0 030 =
020<2:
.0 010 2
Jo0
O
2C
1E
'JEC
1(
ci
1)
762
I\ --
F
763The pathogenesis ofpneumatic tourniquet paralysis in man
100 r
EEa, 50EI-.
0
100
cEi- 5
0
S
-~~~~~~~~~~~0 *- 0
--->80 0 4~a 0_
0 q_--° 0
O 0
I--S2 27T do
5.
75 r -. *.
o Distal to or underneath cuff* Proximnl tn r1 iff
I:~~~~~~~~-VIooIutI)ui
0*--=-000 0
0 0
Sv* *0 o 0 0 v v
o -v
--ooo
-S2
SU
FCU|
N 50
l-
-W
E> 250
> HT
0 10t Tourniquet on
100 200 300Distance from motor point (mm)
Fig 6 Plots of the relative times to initial o
block (upper) and complete conduction bloc)the hypothenar "M" response at various disproximal to the hypothenar motor point. Stipoints proximal to the pneumatic tourniquetdistinguished by the solid circles. The plots j
combination of observations from eight patiThe time to the earliest conduction block an
conduction block was a function of the distcthe point ofstimulation and the motor point
duction block and conduction delay wthe ischaemic length of the nerve betw(of stimulation and the stigmatic electThis was substantiated by the observatime to 50% conduction block in HTwhen the ulnar nerve was stimulated jto an upper limb tourniquet was 20-25when the ulnar nerve was stimulated jto a forearm tourniquet, the time to cblock was longer, namely 30-40 minute
Further evidence to support the irnerve length as a determinant of theduction block was provided by teststimes to equivalent block were compariflexor carpi ulnaris (FCU) motor fibres istimulation proximal to an upper arn
-o-o-o- ....0........
Si- S2FCU0.
0..***
Sl- S2 HT - \
'. I,~-
20 30 40Time (min)
I Fig 7 Upper: Left-schematic illustration of the relative400 500 positions of the recording electrodes (surface) over the
flexor carpi ulnaris (FCU) and hypothenar (HT) musclesconduction in relation to the pneumatic tourniquet about thek (lower) in mid-upper arm and the stimulating electrode; locatedstances just proximal to the tourniquet. Right-plot of theimulation changes in the respective FCU and HT "M" responseare maximum p-p V compared to their initial values at therepresent a various times after tourniquet inflation. Lower: Left-ents. relative positions of the stimulating and recording7d complete electrodes in relation to the tourniquet are the same asrnce between above except for the inclusion of an extra stimulating
electrode (S,) just distal to the tourniquet border zone.
Right-changes in the maximum conduction velocity{as related to m/s in flexor carpi ulnaris (FCU) and hypothenar (HT)een the point motor fibres across the segment S-S2. Note the largertrode(fi degree of conduction block and greater reduction in the
trode (fig 6). maximum conduction velocity in hypothenar motorLtion that the fibres at corresponding times following tourniquetmotor fibres inflation.just proximalminutes but
just proximal Not only was the time to block longer for the shorterorresponding FCU fibres but the reduction in conduction velocitys. was much larger for the HT compared to the FCUnportance of fibres across the tourniquet segment at equivalenttime to con- times after tourniquet inflation. The above evidenceon which the therefore established that the time to the earliested in HT and detection of abnormalities in conduction and thein response to degree of those functional abnormalities were an tourniquet, function of the length of the nerve in the ischaemic
0
Yates, Hurst, Brown
zone. The shorter the length of the nerve tested in theischaemic territory, the longer was the time toconduction block and the less the degree of con-duction delay.The ischaemic zone included both the segment
beneath the tourniquet and the zone distal to thetourniquet. Therefore, it was possible that the samefactor which accounted for the proximal to distalprogression in the degree of abnormality in nervefunction which was observed distal to the tourniquetaccounted also for the even earlier conductionabnormalities at the level of the tourniquet borderzone. With the methods employed, the last represen-ted the longest accumulative length of nerve tested.This led to a fundamental question. If the proximalto distal progression in functional abnormalitiesobserved was a function of ischaemia alone, werethere any abnormalities which could be directlyrelated to mechanical compression?
(4) EVIDENCE FOR MECHANICALCOMPRESSIONTo establish mechanical compression as a factorrequired that abnormalities in nerve function beobserved beneath or across the tourniquet borderzones which could not be explained by the orderlyproximal-to-distal progression in abnormal nervefunction observed in the pure ischaemic zone distalto the tourniquet. Disproportionately larger con-duction delays and block would have to be demon-strated across the distal tourniquet border zonecompared to the more proximal sub-tourniquetsegment. This was indeed evident in some subjectsin both motor and sensory fibres within the first 30minutes (figs 3, 5). It was not, however, possible todemonstrate conduction abnormalities across thetourniquet border zones which were not accom-panied by other parallel, though less severe, conduc-tion changes distal to the tourniquet. This meant thatwhatever mechanical pressure factors were operativeat the level of the tourniquet, these factors werelikely to be overtaken by the general ischaemicchanges. The mechanical pressure factor was perhapsmore obvious in the post-tourniquet release periodwhen, at an early stage, a disproportionate increasein conduction time and block was very obviousacross the distal tourniquet border zones comparedto the more proximal sub-tourniquet and proximaltourniquet segments (fig 3). Even at 1 hour, con-duction delays were commonly present across thetourniquet segment when there had been 100o%recovery distal to the tourniquet. In order to lookfor a distinct pressure factor an attempt was made toseparate the possible ischaemic and pressure factors(fig 8). In this test two tourniquets were employed,one about the mid-forearm and the other about the
| T2 100 O 0
2s X0 2 0 40 50 6
S2 ) 50 \2TITi F7T1 0 20 30 40 50 60
HT) T T2T *1T
Fig 8 Left: The relative positions of the stimulatingelectrodes, S1 at the level of the wrist and S2 at the levelof the elbow compared to the two tourniquets, T1 aboutthe mid-forearm and T2 about the mid-upper arm.Right: Changes in the maximum hypothenar "M"response p-p V compared to the initial pre-tourniquetvalues at various times following tourniquet inflation inresponse to stimulation at the S1 and S2 levels. At time 0T, was inflated to 300 mm Hg. When conduction blockwas clearly established in response to stimulation at S2the upper arm tourniquet was inflated (T2) and shortlythereafter, the forearm tourniquet (T1) was released.The upper arm tourniquet (T2) was finally released at50 min. Note, no reversal in the conduction blockacross the forearm was observed even in the initialperiod.following release of the forearm tourniquet.
mid-upper arm. T1 was inflated first. When a 500%block was evident in ulnar-HT motor fibres acrossthe forearm, T2 was inflated and T1 released. Theprimary question was whether there was anyreversal of the functional abnormalities across theTi when the mechanical compression was releasedthough the nerve remained in the ischaemic zone.If the block in conduction had been primarily theresult of mechanical compression, reversal in theblock would be expected. If, however, the primaryfactor was ischaemia no change in the rate ofprogression in the block would be expected. Therewere mixed results. In some subjects no significantchange in the rate of progression of the conductionabnormalities was observed or if present (fig 8) thetemporary reduction in the rate of progression wasobserved both across the forearm tourniquet seg-ment and distal to the tourniquet. This suggestedthat this partial recovery may have been the result ofrelease into the forearm and hand regions of oxy-genated blood trapped distal to the upper armtourniquet. In other subjects there was a partialreversal of the conduction block observed when Tiwas released. This reversal was never complete.
764
The pathogenesis ofpneumatic tourniquet paralysis inman7
There may, therefore, be a mechanical factor butthis is rapidly overtaken by the ischaemic changes.
Discussion
The primary question in this investigation waswhether functional abnormalities in human nervescompressed by a tourniquet for conventional timesand pressures were the result of ischaemia ormechanical compression or a combination of bothfactors. Ischaemia is the only factor distal to thetourniquet. The issue was, however, complicatedat the level of the tourniquet where nerves wereexposed to both ischaemia and mechanical com-pression. The distal tourniquet border zone wasboth compressed and ischaemic while the proximaltourniquet border zone included a proximal seg-ment which was not ischaemic together with a distalsegment which was both compressed and ischaemicbeneath the tourniquet. Both border zones weresubject to non-uniform pressures and stress patternsbeneath the edges of the tourniquet.5 The worstcombination of ischaemia, mechanical compressionand non-uniform pressure distributions wouldtherefore probably be present across the distaltourniquet border zones.
In the pure ischaemic zone distal to the tourniquet,there was evidence for a proximal to distal pro-gression in the order of onset and severity of con-duction abnormalities in the nerves. This patternhas been reported earlier.6 The proximal to distalprogression may even explain the earliest conduc-tion block and disproportionate conduction delaysobserved across the proximal tourniquet borderzone and across the tourniquet segment as a whole,compared to the zone distal to the tourniquet. Theorder of progression in conduction block distal tothe tourniquet was to be expected on the basis ofother evidence too, on the effect of ischaemia onnerve conduction.7 8 The reverse of the abovepattern however, would be expected if the proba-bility of conduction block per unit length of nervewas uniform along the length of the nerve.9 Theeffect of this would be for the maximum reductionin the "M" response per unit length of the nerve todiminish from distal to proximal; the reverse of thepattern observed here (Appendix A, fig 9). Theevidence therefore suggested that the probability ofconduction per unit length of nerve was not uniformbut higher over the more proximal segments of thenerve. This was substantiated by the dispropor-tionately larger increases in conduction times overthe more proximal segments of the nerve. Thispartially explained why at equivalent times aftertourniquet inflation, the degree of conduction blockwas higher in the longer HT fibres than in the shorter
FCU fibres.The actual basis for the observed proximal to
distal progression in the severity of the conductionabnormalities in the ischaemic nerve has not beenexplained. There was no obvious reason to expectthis progression if all segments of the nerve proximalor distal were exposed to the equivalent degree ofischaemia. Moreover, the opposite would beexpected if there were a proximal to distal fall-offin temperature or if there was any leakage of bloodpast the tourniquet to tissues just beyond the cuff.The latter was unlikely because at the tourniquetpressures employed it was rare to observe any bloodleakage in surgical fields beyond the tourniquet.There are no anatomical or functional characteristicsof nerve known to the authors which properlyexplain this observed proximal to distal conductiorntime gradient. Hence, the primary factor to explainthe abnormalities in conduction at and distal to thetourniquet was probably ischaemia. This wassuggested earlier by Landau.10 It was, however,possible that mechanical pressure effects if theyrequired a sufficiently long enough time to develop,may have been masked by the effects of ischaemia.There were indeed, certain conduction abnormalitiesin the early post-inflation and later release periodswhich pointed to probable local mechanical con-tributions to the observed abnormalities in nervefunction particularly across the border zones of thetourniquet.
Experimental evidence has emphasised the criticalimportance of non-uniform pressure distributions atthe border zone of the tourniquet in the patho-genesis of characteristic structural abnormalities innerve compressed by the tourniquet.4 1112 However,in these experimental preparations, the compressiontimes and pressure levels were substantially abovethose employed in our human investigations. Ofcourse it is hard to compare directly the human andanimal experiments. In view, however, of thecomplete recovery in nerve function in our experi-ments, in most by 1 hour and in all by 24 hours,whatever temporary structural abnormalities mayhave accompanied the compression, these werereversible.
Pneumatic tourniquet injury to peripheral nervemay be more common than is appreciated even atconventionally accepted compression times andpressures. For example, denervation has beendetected in lower limb muscles distal to the tourni-quet in knee operations.'3 Therefore, the tourniquet,even as generally employed has the potential forinjury to peripheral nerves. This risk is enhancedwhen longer compression times or higher pressuresare employed. Higher pressure is a real danger;at our own hospital, calibration of tourniquets in
765
766
common use revealed at least three which registeredpressures 50-100 mm Hg above that which theinstrument itself registered. There is a need forfrequent recalibration of tourniquets and properinstructions of personnel directly responsible fortheir use.The present investigations have revealed that the
primary factor accounting for observed functionalabnormalities in human nerves compressed by thepneumatic tourniquet is ischaemia. From this itmust not be concluded that ischaemia is the primarycause of tourniquet paralysis. Indeed quite theopposite is probably true in so far as ischaemia byitself, even when it lasts much longer than com-pression, does not cause persistent conduction blockand especially the localised conduction delays andconduction block observed across the tourniquetborder zones in tourniquet paralysis.'4 What ismeant rather is that when compression and ischae-mia are combined, as they are in the tourniquetlesion, the functional abnormalities resulting fromischaemia develop sufficiently early at the level of thetourniquet as to largely mask the direct mechanicaleffects. Even so, functional abnormalities attribu-table to pressure alone are detectable. For example,disproportionate abnormalities in nerve function atthe level of the proximal and distal border zones ofthe tourniquet point to the probable contribution ofmechanical compression as the primary factor.Whatever other compressive effects may have beenpresent were lost in the progressive ischaemicconduction block. Therefore at the time of tourni-quet compression it is hard to weigh the relativecontribution of ischaemia and mechanical com-pression to the observed functional abnormalitiesin the nerve. It is clear that whatever the mecha-nisms responsible, the functional abnormalitiesquickly reverse following tourniquet release whenthe compression times are relatively short and thepressures not excessive. Excessive pressure orcompression time may however produce conductionabnormalities localised to the border zones of thecuff which may not promptly reverse and are mostlikely due to mechanical compression.3
APPENDIX ATo predict the degree of conduction block evidentin changes in the "M" response p-pV at increasingdistances from the motor point a theoretical modelhas been constructed. The critical assumptions in thesimple model include: (1) The probability ofconduction block per unit length (internode length)of nerve is constant along the length of the nerve. Letthis probability = p. (2) Each motor unit makes thesame surface voltage contributions to the "M"response voltage. Let that voltage be V. (3) All motor
Yates, Hurst, Brown
fibre conduction velocities are the same. Also let ybe the number of nerve fibres in this model and letthere be n number of internode lengths. The modelmade no adjustment for an increase in temporaldispersion with an increase in distance from themotor point. Experimentally, however, there wasno substantial increase in p-p duration of the "M"response with distance from the motor point beforeor following tourniquet inflation.
QuestionWhat is the degree of conduction block expressed asa percentage reduction in the "M" p-pV at increasingdistances between the point of stimulation and themotor point?
1000
Fig 9 Plot of the percentage of nerve fibres conductingat various numbers of internode lengths away from therecording electrode at "O". For the various plots theprobabilities of conduction block across any oneinternode length are shown.
The number of fibres contributing to the "M"response at various distances from the motor pointmay be predicted by:
If the number of nerve fibres contributing to the"M" response at distance "O" (here the motorpoint) = y and the voltage generated per nervefibre equals V, then the "M" response voltageat "O" = y x v. Now let p be the probability ofconduction block in each fibre per unit length(here considered to be one internode length) thenat the 1st internode the number of fibres conduct-ing = y - (y x p) or y(l - p).At the 2nd internode-number of fibres con-ducting = y(1 - p) - y(1 - p) x por y(I - p)2At the 3rd internode-number of fibres con-ducting = y(1 - p)2 - y(l - p)2 X por y(1 - p)3
The pathogenesis ofpneumatic tourniquet paralysis in man
Finally, at the nth internode-number of nervefibres conducting = y(l - p)'1; therefore, thevoltage at distance n number of internodes =y(l - p)n x V.This model predicts that the maximum reduction
in p-pV per unit length of nerve will be evident overthe distal segments-there being a reduction in the"apparent" degree of block with increasing distancefrom the motor point. This progression stood indirect contrast to the proximal to distal progressionobserved. Mechanical compression could explain theearliest conduction block at the level of the tourni-quet. However, this seemed unlikely in view of theprogression distal to the tourniquet which wasobserved to follow soon after the earlier changes atthe level of the tourniquet. It is possible that theprobability of conduction block per unit length ofnerve was not uniform but increased from distal toproximal. This fitted best with the evidence.
References
Rudge P. Tourniquet paralysis with prolonged con-duction block. JBone Joint Surg (Br) 1974;56:715-20.
2 Sunderland S. Nerves and Nerve Injuries. 2nd ed.Edinburgh: Churchill Livingstone, 1978.
3Bolton CF, McFarlane RM. Human pneumatictourniquet paralysis. Neurology (Minneap.) 1978;28:787-93.
4Ochoa J, Fowler TJ, Gilliatt RW. Anatomical changesin peripheral nerves compressed by a pneumatic
9 Waxman SG, Brill MH, Geschwind N, Sabin TD,Lettvin JY. Probability of conduction deficit asrelated to fiber length in random distribution modelsof peripheral neuropathies. J Neurol Sci 1976;29:39-53.
10 Landau WM. Mechanism of nerve block. Nature1972;237:224.
Grundfest H. Effects of hydrostatic pressures upon theexcitability, the recovery and the potential sequenceof frog nerves. Cold Spring Harbour Sym7p Biol 1936;4:179-87.
12 Bentley FH, Schlapp W. The effect of pressure onconduction in peripheral nerve. J Physiol (Lond)1943 ;102 :72-82.
13 Weingarden SI, Louis DL, Waylonis GW. Electro-myographic changes in post meniscectomy patients:role of the pneumatic tourniquet. JAMA 1979;241:1248-50.
14 Hess K, Eames RA, Darveniza P, Gilliatt RW. Acuteischaemic neuropathy in the rabbit. J Neurol Sci1979 ;44 :19-43.
