Annals of the Royal College ofSurgeons of England (1986) vol. 68
Doppier waveform analysis in the management
of lower limb arterial disease
W BRUCE CAMPBELL MS FRCS*
Research Fellow and Honorary Senior Registrar Bristol Royal Infirmary
Key words: ARTE.RIIES; ARTLRIR()S(CI.IR,RoSIS; I)IAGNOSIS, (:ONPJTEIR ASSISTEA): UILTRASONI(:S; \V\S(UJILAR D)ISlASIA
Arterial disease changes the shape ofDoppler ultrasound waveforms
recorded noninvasivelyfrom arteries in the lower limbs. These changes
can be described numerically by computer analysis ofwaveforms, and
techniques currently in use are pulsatility index, Laplace transform
and principal component analysis.
These waveform analysis methods allow assessment of aortoiliac
disease from Doppler recordings at the femoral artery. In addition,
thefemorodistal segment can be evaluated by comparingfemoral and
distal waveforms, even in the presence of more proximal arterial
disease. Other possible applications for waveform analysis include
noninvasivefollow-up ofangioplasty or bypass grafts and the detec-
tion ofearly, presymptomatic arterial disease. Experience in recording
Doppler waveforms is important if variability is to be minimised.
Currently, these methods allow exclusion ofaortoiliac disease prior
tofemorodistal grafting but their other potential roles requirefurther
In 1959 Satomura described the use of a Doppler ultra-
sound probe for detecting blood flow signals from arteries
(1) and suggested that this technique might be used in the
diagnosis of atherosclerosis. During the 1960's the applica-
tions of the Doppler ultrasound probe in vascular practice
were more clearly defined, in particular, confirmation of
vessel or graft patency and the measurement of lower limb
systolic pressure (2). All these could be achieved simply by
listening to the sound generated by the Doppler system.
However, records ofwaveform shape also became available
and directional systems were developed which could sepa-
rate forward and reverse flow during the various phases of
the cardiac cycle (3) (Fig. la).
It was observed that arterial disease resulted in charac-
teristic changes in the shape of the Doppler waveform,
which could be recognised by simple inspection of the trace
(4) and described, to some extent, by reference to specific
waveform features; loss of the early diastolic reverse flow of
amplitude etc. (Fig. lb & c). These changes can be used to
diagnose disease in vessels proximal to the site of recording
either by simple observation (5) or by measurement of
specific parameters from the waveform (6).
Recognition of progressive changes in ultrasound wave-
form shape invited development of an objective method by
which they could be measured. The first objective techni-
que of waveform analysis was described by Gosling and his
colleagues who devised a parameter called pulsatility index
*Present Appointment and correspondence to: Clinical Lecturer
in Surgery, Nuffield Department of Surgery, John Radcliffe
Hospital, Oxford OX3 9DU
(PI) which was obtained by dividing the maximum vertical
excursion of the waveform by the mean value throughout
the cardiac cycle. The value of PI was demonstrated to fall
1 The shape of Doppler arterial waveforms in health and
disease. Doppler shift frequency is on the vertical axis (pro-
portional to blood velocity), and time is on the horizontal axis.
(a) shows a normal waveform, with rapid systolic upslope and
downward slope. There is a phase of reverse flow in early
diastole, followed by a phase of forward diastolic flow.
(b) This waveform, distal to a severe stenosis shows prolongation
of the systolic downward slope with less marked prolongation
of the upslope of the wave. The early diastolic reverse flow has
been lost-this is a characteristic finding distal to a stenosis.
(c) This waveform, distal to a complete occlusion, is recorded at
higher gain than (a) or (b). Blood is bypassing the upstream
occlusion through collaterals and the wave is very damped.
Forward flow throughout diastolie may be seen in severely
ischaemic limbs with dilated peripheral vessels.
W. Bruce Campbell
when the ultrasound waveform became damped distal to an
arterial stenosis or occlusion (7, 8). Despite certain prob-
lems in its use (9, 10), PI has stood the test of time (11) and
remains a viable method of waveform analysis with which
other newer techniques have been compared.
The two most important new methods of Doppler wave-
form analysis to have emerged in recent years are the
Laplace transform method (12-14) and principal compo-
(15, 16). Both are complex mathematical
techniques and are performed by computer. The Laplace
transform method produces coefficients related to waveform
damping (b), arterial elasticity (wo), and peripheral resist-
ance (y). Principal component analysis yields two principal
components whose values can be plotted graphically to
produce a two-dimensional display or, alternatively, the
first principal component (PC1) alone can be used as a
measure of arterial waveform damping.
Pulsatility index (PI), Laplace transform analysis and
favoured by different groups of investigators. Used in isola-
tion or side by side, they have now been extensively evalu-
ated in the management of lower limb arterial disease.
have each been
The diagnosis of aortoiliac stenosis
The first major clinical area in which Doppler waveform
analysis appeared to offer helpful information was the
diagnosis of aortoiliac stenosis. In the case ofa tight stenosis
no diagnostic problem exists. The femoral pulse is weak,
there may be a bruit and disease is confirmed by arteri-
ography. If the patient's symptoms merit treatment the
need to tackle a severe stenosis by bypass grafting or
angioplasty is obvious.
The problem arises in the patient who has an obvious
occlusion in the femoropopliteal segment which would seem
suitable for a femoropopliteal bypass graft, but in whom
there is uncertainty about the state of the proximal arteries.
If there is a stenosis in the aorta or iliac vessels which
escapes detection, this predisposes to failure of any more
distal graft (17). Some patients with less severe stenosis of
the iliac arteries may have normal femoral pulses and, even
after arteriography, there may be problems in assessing the
aortoiliac segment (18). Disease adjacent to the anterior or
posterior wall of the vessel may not show up as narrowing
on an ordinary antero-posterior view. Even when narrow-
ing is apparent it may be difficult to be sure how severe it is
from a haemodynamic point of view. Biprojectional arterio-
grams help to minimise this problem but may still leave
room for doubt. Detection of less severe stenoses may
become increasingly important with greater availability of
transluminal angioplasty. This provides a simple method
for treatment of these lesions providing optimum inflow for
any more distal graft.
Analysis of common femoral artery waveforms supple-
ments clinical assessment of the aortoiliac segment (19) and
can provide prognostic information on the fate of femoro-
popliteal bypass grafts, by detecting proximal disease (17).
When comparing the different waveform analysis methods
for this pupose, one problem is choosing an appropriate
'gold standard' for reference. Although arteriograms are
imperfect, they are available for large numbers of patients,
and PI (9, 19, 20), Laplace 6 (9, 19) and PCA (16) have all
been shown to have some correlation with the arterio-
graphic severity of aortoiliac stenoses.
measurements provide a more accurate reflection of the
degree ofproximal disease and the three methods ofanalys-
ing common femoral artery waveforms have also been
demonstrated to correlate with pressure gradients (11, 16,
19). Broadly speaking, all methods can detect stenoses of
greater than 50% diameter reduction. The argument cen-
tres on which detects lesser stenoses most reliably with
conflicting evidence from clinical studies favouring the
Laplace transform method (19) and PCA (16). One source
of inaccuracy may be the presence of occlusive disease in
the superficial femoral artery, which can produce falsely
elevated PI results (9, 10). Changes in distal resistance have
also been claimed to affect Laplace 6 values (15) although
this is controversial (21).
Despite these problems it seems that significant aortoiliac
disease can be virtually excluded by a normal waveform at
the common femoral artery. An abnormal waveform pro-
vides grounds for caution and it is reasonable in these cases
to measure aortofemoral pressure gradients (with flow en-
hancement, for example by using papaverine) before im-
planting a femorodistal graft.
Assessment of the femorodistal segment
The section above has dealt with the problems of detecting
aortoiliac disease before proceeding to a distal bypass graft.
There is also a case for assessing the haemodynamic state of
the distal vessels prior to a proximal reconstruction. In
multisegmental disease a proximal reconstruction should
generally be the first step and this usually gives relief of
symptoms. However, a proportion of patients continue to
get symptoms and, furthermore, some cases of severe
ischaemia may be most successfully dealt with by combined
proximal and distal grafting. It is currently difficult to select
these cases with confidence. Analysis of waveforms from the
common femoral artery and from more distal sites offers a
potential method for assessing the intervening arterial seg-
Pulsatility index has been used in this way. The value
obtained from dividing proximal PI by distal PI, the PI
damping factor, is normally greater than one. If there is an
intervening occlusion the distal PI falls and the resulting PI
damping factor rises (8).
The Laplace transform technique has also been applied
to the femorodistal segment. The Laplace coefficient wo
(related to arterial elasticity) has a higher value in wave-
forms from ankle arteries than from common femoral artery
waveforms in normal limbs, but in the presence of an
intervening occlusion the distal value falls. This 'wo gra-
dient' (femoral Laplace wo divided by distal Laplace w0)
has been compared with PI damping factor and its accu-
racy in detecting femoropopliteal occlusions shown to be
similar, with rather less false positive results (22). The most
important aspect of waveform analysis in the femorodistal
segment is its ability to assess this part of the arterial tree
even in the presence ofaortoiliac disease. This should allow
selection of patients in special need of combined proximal
and distal grafting procedures. Doppler also detects patent
distal vessels required for grafting, although the final
decision regarding a distal graft is usually made on the basis
of detailed arteriography. The clinical application ofwave-
form analysis in this area requires further work.
Follow-up ofpercutaneous transluminal angioplasty
and ofbypass grafts.
The fact that waveform analysis can assess one arterial
segment separately from another has led to its evaluation in
the follow-up of surgical reconstruction and angioplasty.
The technique is certainly of some value in assessing iliac
angioplasty, by analysis of common femoral artery wave-
form, although any haematoma may interfere with record-
ings, both early on and in the longer term (23) and scarring
at the groin following bypass surgery can also cause
problems. In addition, the advantage ofwaveform analysis
in assessing the femorodistal segment separately from the
proximal arteries is reduced by its lack of sensitivity com-
pared with simple ankle pressure measurements (21,23).
The detection ofpresymptomatic arterial disease
A noninvasive method of detecting early, presymptomatic
arterial disease using the relatively accessible lower limb
vessels would offer a number of interesting possibilities.
Firstly, it would facilitate the study of the natural history of
Doppler waveform analysis in the management of lower limb arterial disease
atheroma and the influence of suspected risk factors on its
progress. Secondly, an easily repeatable method ofmeasur-
ing atheromatous changes might allow monitoring of any
non-surgical therapy for atheroma, such as reduction of
blood lipid levels. Thirdly, a simple non-invasive method of
detecting individuals with presymptomatic disease might
enable them to be brought under medical supervision at an
early stage with a view to measures aimed at reducing their
risk of serious complications of arterial disease, such as
myocardial infarction or stroke.
The fact that Laplace waveform analysis can detect
minor and subtle changes in Doppler waveform shape
suggested that this method might have a role in the detec-
tion of early disease. In order to investigate this possibility,
subjects were selected from over two thousand men in
Bristol, involved in a longitudinal study of risk factors in
cardiovascular disease (24). At their intitial screening ex-
amination all had undergone recording of Doppler wave-
forms at both ankles for Laplace analysis. Three years later
ninety were selected for more detailed study-49 with high
values of Laplace 6 or wo (in the top tenth percentile) and
41 with lower values. In the group with high initial wave-
form analysis values, 11 of 49 (22%) were found to have
signs of arterial disease while among those with lower
values only 2 of 41 (5%) had evidence of arterial disease
(P<0.05 by X2 analysis). This finding supported
hypothesis that Laplace analysis of waveforms obtained
from arteries about the ankle could identify a group with an
increased incidence of arterial disease.
The study also revealed problems with the method.
There was considerable variability of results in individual
subjects between the intitial screening and examination and
that 3 years later. Also, there were marked differences in
waveform analysis results obtained by two different oper-
ators in many of the subjects. Finally, the study failed to
identify one of the two Laplace coefficients examined (6 or
wo) as being consistently successful in identifying patients
with arterial disease. Despite these criticisms, this study
drew attention to the possibility ofdetecting arterial disease
in its early stages by simple noninvasive means, which
seems a prerequisite to effective early treatment of athero-
sclerosis on a wide scale.
Variability and Reproducibility
Recording Doppler waveforms for analysis requires skill
and experience, and increasing practice enhances repro-
ducibility (25). Incorrect use of the Doppler probe or in-
accurate positioning can produce seriously aberrant results
and this may be a particular problem when there are large
plaques ofatheroma in the region ofrecording. Physiologic-
al changes in the patient, especially vasodilation and vaso-
contriction, influence waveform shape, particularly at more
distal sites in the limb. Finally, different types of Doppler
processing equipment may have a minor effect on repro-
ducibility (20) although this seems very much less impor-
tant than variability attributable to the user (25).
Analysis ofDoppler waveforms from the lower limb arteries
remains an area of fascinating promise but limited fulfil-
ment in clinical practice.
Currently the most important application of the tech-
nique is in excluding significant aortoiliac disease prior to
distal bypass grafting. In clinically and arteriographically
doubtful cases, however, abnormal femoral waveforms
should perhaps be supplemented by direct measurement of
pressure gradients before deciding on a proximal graft or
In the femorodistal segment, comparison of waveform
analysis values from the femoral and distal arteries help to
evaluate the severity of disease, even when the aortoiliac
segment is stenosed or occluded. This may allow selection
of patients for combined proximal and distal grafting but
the method needs to be tested prospectively in appropriate
Waveform analysis may provide a useful additional
method for follow-up after angioplasty or bypass grafting.
In particular femoral artery recordings can provide in-
formation about the iliac segment following angioplasty. In
the femorodistal segment variability reduces the value of
the method compared with less specific measurements such
as ankle systolic pressure index.
Finally, there may be a role for waveform analysis in
assessing early arterial disease. A major problem in this
area is reproducibility, which may be affected by operator
experience, physiological changes and the Doppler system
used. The success ofthis technique for any purpose depends
on skilled use ofthe equipment since there is a tendency for
the less practised operator to obtain 'abnormal' waveforms.
It is a pleasure to acknowledge the help and support I received
during my studies on Doppler waveform analysis from Mr R N
Baird, Consultant Surgeon, Bristol Royal Infirmary, and from
members of the Department of Medical Physics-Professor J P
Woodcock, Dr R Skidmore and Miss S E A Cole. I am grateful to
Dr D H Evans for providing the programme for principal compo-
nent analysis. I also thank Mr W S Paige for his assistance with
statistical analyses and Imperial Tobacco Limited for their gener-
ous financial support.
1 Satomura S. Study ofthe flow patterns in peripheral arteries by
ultrasonics. J Acoust Soc Japan 1959; 15:151-8.
2 Yao ST, Hobbs JT, Irvine WT. Pulse examination by an
ultrasonic method. Br MedJ 1968;4:555-7.
3 Berstein EF, Murphy AE, Shea MA, Housman LB. Ex-
perimental and clinical experience with transcutaneous Dop-
pler ultrasonic flowmeters. Arch Surg 1970;101:21-5.
4 Nayman J. The use of the ultrasonic flowmeter in peripheral
vascular disease. Aust NZ J Surg 1974;44:157-67.
5 Persson AV, Gibbons G, Griffey S. Noninvasive evaluation of
the aortoiliac segment. J Cardiovas Surg 198 1;22:539-42.
6 Dilley RB, Fronek A. Quantitative velocity measurements in
arterial disease of the lower extremities. In: Berstein EF ed.
Noninvasive diagnostic techniques in vascular disease. Saint
Louis: C V Mosby 1978. pp 294-303.
7 Gosling RG, Dunbar G, King DH, Newman DL, Side CD,
Woodcdck JP, Fitzgerald DE, Keates JS, MacMillan D. The
quantitative analysis of peripheral arterial disease by a non-
intrusive ultrasonic technique. Angiology 1971;22:52-5.
8 WoodcockJP, Gosling RG, Fitzgerald DE. A new non-invasive
technique for assessment of superficial femoral artery obstruc-
tion. Br J Surg 1972;59:226-31.
9 Baird RN, Bird DR, Clifford PC, Lusby RJ, Skidmore R,
Woodcock JP. Upstream stenosis. Its diagnosis by Doppler
signals from the femoral artery. Arch Surg 1980; 115:1316-22.
10 Hamilton WAP, Fulton TJ, Gay P, Stevens AL, Graham JC,
Roberts VC, Cotton LT. Locating aortofemoral disease by
measuring flow velocity in the femoral artery using Doppler
ultrasound-a comparative study. Vascular Diagnosis and
11 Demorais D,Johnston KW. Assessment ofaortoiliac disease by
non-invasive quantitative Doppler waveform analysis. Br J
Surg 1981 ;68:789-92.
12 Skidmore R, Woodcock JP. Physiological interpretation of
Doppler-shift waveforms. I. Theoretical considerations. Ultra-
sound Med Biol 1980;6:7-10.
13 Skidmore R, Woodcock JP. Physiological interpretation of
Doppler-shift waveforms: II. Validation of the Laplace trans-
form method for characterisation of the common femoral
14 Skidmore R, Woodcock JP, Wells PNT, Bird D, Baird RN.
Physiological interpretation of Doppler-shift waveforms: III.
Clinical results. Ultrasound Med Biol 1980;6:227-31.
15 Evans DH, Macpherson DS, Bentley 5, Asher MJ, Bell PRF.
The effect of proximal stenosis on Doppler waveforms: a com-
parison of three methods of waveform analysis in an animal
model. Clin Phys Physiol Meas 1981;2:l7-25.
W. Bruce Campbell
16 Macpherson DS, Evans DH, Bell PRF. Common femoral
artery Doppler waveforms: a comparison of three methods of
objective analysis in patients with vascular disease. BrJ Surg
17 Charlesworth D, Harris PL, Cave FD, Taylor L. Undetected
aortoiliac insufficiency: a reason for early failure of saphenous
vein bypass grafts for obstruction of the superficial femoral
artery. BrJ Surg 1975;62:567-70.
18 Bruins Slot H, Strijbosch L, Greep JM. Interobserver varia-
bility in single-plane aortography. Surgery 1981;90:497-503.
19 Campbell WB, Cole SEA, Skidmore R, Baird RN. The clini-
cian and the vascular laboratory in the diagnosis of aorto-iliac
stenosis. BrJ Surg 1984;71:302-6.
20 Johnston KW, Maruzzo BC, Cobbold RSC. Doppler methods
for quantitative measurement and localisation of peripheral
arterial occlusive disease by analysis of the blood flow velocity
waveform. Ultrasound Med Biol 1978;4:209-23.
21 Campbell WB. Laplace transform analysis of Doppler wave-
forms in lower limb arterial disease. MS Thesis. University of
22 Campbell WB, Cole SEA, Evans JM, Skidmore R, Woodcock
J, Baird RN. Physiological interpretation of Doppler shift
waveforms: the femorodistal segment in combined disease.
Ultrasound Med Biol 1983;9:265-9.
23 Campbell WB, Cole SEA, Jeans WD, Baird RN. Percutaneous
transluminal angioplasty for lower limb ischaemia. Br J Surg
24 Campbell WB, Skidmore R, Woodcock JP, Baird RN. Detec-
tion of early arterial disease: a study using Doppler waveform
analysis. Cardiovasc Res 1985;19:206-1 1.
25 Campbell WB, Skidmore R, Baird RN. Variability and repro-
ductibility of arterial Doppler waveforms. Ultrasound Med