The concept of a clinical diagnostic vascular laboratory was first formulated two decades ago. Technological advances made in the 1960s allowed methods originally developed for research to be applied to the non-invasive evaluation of blood flow and blood pressure in clinical settings. The earliest such techniques involved indirect methods for evaluating cerebrovascular and peripheral vascular insufficiency. Rapid developments in the technology over the following 20 years have produced both indirect and direct methods of evaluation which provide both anatomical and physiological descriptions of vascular pathology. This chapter reviews the current technology and methods used in the diagnostic vascular laboratory and their various applications in the clinical evaluation and treatment of cerebrovascular and peripheral vascular disease. Footnote 1

Stroke represents the third leading cause of death in the United States today. Since extracranial carotid atherosclerotic disease is one of the major causes of stroke, its diagnosis and treatment is an important issue. Patients at risk for stroke include those who have suffered prior stroke or transient ischaemic attacks, and perhaps those with asymptomatic carotid bruits. Non-invasive studies should be performed on all such high-risk patients to identify those who might benefit from medical treatment or carotid endarterectomy. The studies can also be used to monitor the natural progression of carotid disease in patients with mild to moderate carotid stenosis in whom initial surgical therapy is not indicated, and to rule out postoperative restenosis in patients who have undergone carotid endarterectomy. Most vascular surgeons advocate studies 1, 3, and 12 months after surgery, and annually thereafter.

Non-invasive studies of carotid vascular function fall into the two main categories of direct and indirect tests: direct methods enable the carotid bifurcation itself to be examined while the indirect methods assess the effects of carotid stenosis on the distal circulation (such as ophthalmic artery pressure). Twenty years ago the only non-invasive tests available were indirect methods. In the past 10 years rapid progress in the development of Doppler and pulsed echo ultrasound technology has increased the use of direct methods. Several specialists have suggested that duplex scanning may be the only non-invasive test needed in the evaluation of carotid atherosclerotic disease. This approach, however, fails to take into account the valuable information regarding cerebral collateral flow that the indirect tests provide. Thus many laboratories use indirect tests such as ocular pneumoplethysmography as adjuncts to direct Doppler methods (Table 1) 172.

Ocular pneumoplethysmography was initially designed as a non-invasive method for the determination of carotid stump pressure when applied in conjunction with common carotid test compression. The test then became popular as a non-invasive method of identifying haemodynamically significant carotid stenoses.

This technique uses suction ophthalmodynamometry for indirect measurement of the ophthalmic artery pressure. Small suction-cups applied to the sclera of each eye are capable of detecting arterial pulsations. The application of 300 mmHg suction pressure (500 mmHg in hypertensive patients) to the eye cups by an electric vacuum pump distorts the globe and increases the intraocular pressure. When the intraocular pressure increases above the ophthalmic artery pressure the arterial pulsations in the eye are abolished. The suction is then automatically reduced over a period of 30 s, while continuous strip chart recordings of the suction and the pulsations in the eye are made (Fig. 1) 213. The systolic ophthalmic artery pressure in each eye is determined as the point at which pulsations first reappear. The pressure values recorded on the strip chart are actually measurements of the amount of suction: these values are converted to intraocular pressures using an analogue method.

The presence of haemodynamically significant carotid stenosis is indicated by differences in the ophthalmic artery pressures between the two eyes, the ratio of the ophthalmic artery to the brachial artery pressure, and the pulse amplitude. The test is considered abnormal if there is a difference of 5 mmHg or more between the two eyes; a difference of 1 to 4 mmHg and a ratio of less than 0.66 between the pressure in the ophthalmic artery and that in the brachial artery, a ratio of less than 0.60 between the pressure in the ophthalmic and brachial arteries; or a difference in amplitude of the first pulse of at least 2 mm (this last criterion is applied only to hypertensive patients, in whom the suction pressure does not obliterate the pulse). The accuracy with which haemodynamically significant carotid stenosis can be identified using these criteria is approximately 93 per cent. Ocular pneumoplethysmography is particularly valuable in patients in whom it is difficult to use direct tests; these include those with a densely calcified or highly tortuous carotid, those in whom the vessel has a high bifurcation, and immediately following carotid endarterectomy when the surgical incision prevents application of the Doppler probe. The study should not be performed in patients with glaucoma, lens implant, recent trauma, or those taking coumadin.

One of the most valuable applications of ocular pneumoplethysmography is as an adjunct to carotid duplex studies, to provide information regarding the status of the collateral cerebral circulation. The exact risk of stroke following a carotid arterial occlusion is difficult to predict and largely depends on the presence or absence of collateral flow, which may be extracranial, extracranial–intracranial, or intracranial. Extracranial collaterals may exist between the external carotid and subclavian or vertebral arteries; extracranial–intracranial collateral flow may occur between branches of the facial artery and the ophthalmic artery; intracranial collateral flow may involve the leptomeningeal vessels or the circle of Willis.

The adequacy of these collaterals can be assessed with ocular pneumoplethysmography since the ophthalmic artery pressure is a reflection of both ipsilateral carotid flow and any collateral flow. Several studies have shown an increased risk of stroke in patients with abnormal results of ocular pneumoplethysmography compared to patients with carotid stenosis but a normal test. The combination of ocular pneumoplethysmography and ipsilateral common carotid artery compression provides a measure of the collateral ophthalmic artery pressure. Patients in whom this is less than 60 mmHg are at high risk for stroke should the carotid stenosis progress to occlusion. This test must be performed with extreme caution since the pressure applied to the common carotid artery could cause embolic stroke.

Direct methods examine the carotid artery itself. The Doppler shift principle is used to identify abnormal velocities and flow patterns within the carotid artery, and B mode ultrasonography is used to identify the presence and characteristics of atherosclerotic plaque at the carotid bifurcation. Two types of ultrasonic Doppler velocity metering systems are currently popular: continuous-wave Doppler and range-gated pulsed Doppler.

Continuous-wave Doppler examination of the carotid artery is performed using a hand-held Doppler probe which contains two piezoelectric crystals, one for continuous transmission (usually at a frequency of 5 MHz) and the other for reception (Fig. 2) 214. The probe is ‘coupled’ to the skin with an acoustic gel and held at an angle of 45 to 60° to the skin overlying the region of the common carotid artery in the area medial to the sternocleidomastoid muscle. The probe is advanced distally along the artery, guided by the audible signal and by the spectral frequency display on a video monitor (Fig. 3) 215. The internal and external carotid arteries have unique Doppler shift frequency patterns: the internal carotid has high flow (and thus elevated frequency) in diastole due to the low resistance of the cerebral circulation. The external carotid, on the other hand, has low or almost no flow (Doppler shift frequency approaches 0 kHz) during diastole due to the normally high resistance of the external carotid system (Fig. 4) 216. Because the continuous mode of transmission does not provide any spatial resolution, these unique Doppler patterns identify which artery is being examined. Pathological alterations in flow, such as increased distal internal carotid resistance due to disease, are a source of potential confusion when making this identification.

The audible Doppler signal provides a great deal of information regarding the presence of abnormal flow, but this is very subjective and does not result in a permanent record of the study. More sophisticated methods for analysis of the Doppler shift frequency data have therefore been developed: the most commonly used involves real-time sound spectrum analysis using fast Fourier transform to produce a graphic display of the Doppler spectrum obtained from insonation of the carotid artery. In the display, Doppler shift frequency (which corresponds to blood flow velocity) appears on the vertical axis and time is on the horizontal axis; the amplitude of any specific frequency is indicated by a continuous grey scale. The intensity of the grey scale reflects the number of red blood cells travelling at the velocity which produces that particular Doppler shift frequency. The vertical axis in the spectral display can be expressed as frequency (kHz) or as velocity (cm/s). If a continuous-wave Doppler probe transmitting at 5 mHz is directed at an incident angle of 60° to the carotid artery, a Doppler shift frequency of 1 kHz equals a blood flow velocity of 30 cm/s.

Carotid artery stenosis has two effects on blood flow which can be evaluated with continuous-wave Doppler and spectral analysis: turbulent flow and an increase in the flow velocity. As the degree of stenosis increases, the velocity of flow in the narrowed segment increases, as does the Doppler shift frequency. Flows immediately distal to the stenosis becomes increasingly turbulent, resulting in an increase in the range of frequencies represented in the spectral waveform (special broadening). The continuous nature of the ultrasound transmission used in the continuous-wave technique means that the entire cross-section of the carotid artery is insonated. Since normal blood flow is laminar, with blood travelling along the arterial wall flowing at a slower speed than that in the central portion of the artery, there is a certain amount of ‘spectral broadening’ in an entirely normal vessel, which makes spectral broadening more difficult to discern.

A variety of quantitative analytical methods may be applied to the frequency spectrum data in order to define more precisely the severity of disease in the carotid artery. One popular and fairly simple method of such analysis involves correlating the peak Doppler frequency with the degree of luminal narrowing. It is possible to define four ranges of severity of carotid stenosis on the basis of the peak Doppler shift frequency occurring during systole (Table 2) 173.

The normal systolic peak frequency (f&subm;&suba;&subx;) is 0.5 to 2.5 kHz at the origin of the internal carotid and 2 to 3 kHz in the distal internal carotid artery. The lower f&subm;&suba;&subx; in the carotid bulb is due to its greater diameter. Frequencies of less than 5 kHz are usually seen in mild disease, while f&subm;&suba;&subx; in the 5- to 8-kHz range indicates moderate disease and represents a 30 to 50 per cent reduction in the diameter of the carotid lumen. When f&subm;&suba;&subx; is in the range 8 to 12 the stenosis is in the severe range, with a 50 to 70 per cent reduction in diameter. A peak frequency greater than 12 kHz places the lesion in the category of critical stenosis with a reduction of 70 per cent in diameter and 90 per cent in lumen area (Fig. 5) 217. Such a critical stenosis produces a drop in pressure and flow across the stenosis, due to the severe degree of narrowing.

Another parameter frequently used to assess the spectral analysis is the carotid index which is a ratio of the peak systolic frequency in the internal carotid artery to that in the common carotid artery. The f&subm;&suba;&subx; and the carotid index correlates with the residual lumen diameter of the carotid artery: f&subm;&suba;&subx; 7.5 and a carotid index above 3.8 are indicative of a residual lumen of less than 2 mm. An f&subm;&suba;&subx; above 14 kHz and a carotid index of more than 7 almost always indicates a residual lumen diameter of less than 1 mm.

When the severity of a carotid stenosis progresses such that the cross-sectional area of the artery is reduced by more than 90 per cent, or a residual lumen diameter less than 1 mm (preocclusive range), the blood flow may approach zero as the stenosis approaches occlusion. This may result in a falsely low or absent Doppler shift and a misdiagnosis of no stenosis or an incorrect diagnosis of occlusion, respectively. This potential for error must be kept in mind since either misdiagnosis would lead to surgery not being considered, leaving the patient at high risk for stroke (Fig. 6(a,b)) 218.

A true carotid occlusion is indicated by the absence of a Doppler shift in the internal carotid artery, a decrease in peak velocity, and decreased or absent diastolic flow component in the common carotid artery (i.e. an increased pulsatility). The waveforms obtained from the common carotid artery are also asymmetrical, compared to those of the contralateral carotid artery, and the blood flow in the contralateral normal carotid may be elevated (Fig. 7(a,b)) 219.

Continuous-waver Doppler is a very effective and efficient method of direct non-invasive assessment of the carotid artery. However its lack of spatial resolution may result in occasional diagnostic errors. It may be difficult to ascertain the angle of incidence between the probe and the flow velocity vector in the carotid artery, which can produce errors in the Doppler shift frequency, especially when insonating very tortuous vessels. If the zone of high flow velocity is restricted to a small area in the artery it may be impossible to locate with the continuous-wave sound beam. Other sources of error include anatomical variations, insonating the wrong branch vessel, and the presence of venous flow. Some of these problems may be solved by visualizing the carotid bifurcation with B-mode ultrasound and using pulsed Doppler to allow more precise insonation of the vessel for velocity measurements.

Duplex scanning has become the most common non-invasive method in the examination of the carotid arteries. This technique combines the imaging capability of B-mode ultrasound and the velocimetric capabilities of range-gated (pulsed) Doppler in a single instrument. B-mode ultrasound produces a standard two-dimensional image of the carotid artery, using the sound wave reflection that occurs between tissues of different acoustic densities (Fig. 8) 220. This image provides a ‘map’ which may be used to direct the insonation of the vessel. The pulsed Doppler probe consists of a single crystal which serves both transmitting and receiving functions. By varying the time between emission and reception of the ultrasound signals, flow at different depths within the tissue can be investigated. The site which is actually studied for the Doppler shift effect has a small volume (1×1×2 mm). Unlike continuous-wave Doppler, which provides a spectral array that represents the entire cross-section of the artery, pulsed (or range-gated) Doppler studies therefore provide a spectral pattern obtained from a small sample volume (gate) within the artery.

Rapid technological advances over the last decade have produced a wide variety of duplex scanners with different computer software capabilities and different transducer head designs. The transducers may be mechanical sector scanners or electrical linear or phased array transducers. The most versatile arrangement uses separate transducers for imaging and for Doppler examination, allowing the acquisition of the range-gated Doppler signal to be interpolated among the pulse-echo imaging signals. This provides continuous Doppler and real-time imaging without interruption of either modality.

The duplex scan is performed by placing the transducer on the neck overlying the region of the carotid bifurcation. A real-time B-mode image of the carotid artery is displayed on the monitor screen. The position of the Doppler sample volume (gate) is then adjusted by reference to a line that appears on the B-mode image indicating the angle of incidence between the Doppler beam and the carotid. The actual location of the gate is indicated on the monitor as a small rectangle. The scanner is equipped with a printer that can provide a hard copy of both the B-mode image and the Doppler spectral waveform (obtained by fast Fourier transform) on the same print. Depending upon the computer software package of the scanner, the Doppler signal analysis may include maximum and mean frequency waveforms, computed pulsatility index, and other analyses of the waveform. The spectral waveform is displayed on the same format as that described for continuous-wave Doppler. The grey scale which indicates frequency amplitudes may be in black and white or may be colour coded (Fig. 9) 221.

Since the pulse method provides Doppler shift frequency data from a small volume within the artery, as opposed to the entire diameter of the vessel sampled by continuous-wave Doppler, the range of frequencies displayed in a normal artery is less when using the gated technique. Pulsed Doppler is thus more sensitive for detecting turbulent flow distal to a stenosis. The spectral broadening produced by the turbulent flow appears as a filling in of the spectral window on the waveform (Fig. 10) 222.

The degree of carotid stenosis is assessed mainly on the basis of the spectral analysis, particularly with respect to peak systolic frequency, spectral broadening, end-diastolic frequency, and contour of the waveform. The vertical axis of the spectral plot may be expressed as either frequency or velocity. Various authors have described categories of severity of disease based on quantitation of the velocity or frequency shifts (Table 3) 174.

Duplex scanning eliminates some of the difficulties inherent in continuous-wave Doppler examination by providing an image of the vessel which allows precise insonation of the vessel. However there are several potential sources of error that may affect duplex velocimetry. Low cardiac output or an innominate or proximal common carotid stenosis may decrease flow velocities, thus causing an underestimation of existing bifurcation disease. As with continuous-wave Doppler, if the stenosis is so severe that there is minimal flow, the area of flow may be missed when positioning the gate, leading to a misdiagnosis of carotid occlusion. Despite these limitations, duplex scanning represents a powerful method of direct testing.

Standard duplex imaging provides Doppler shift data from single small volumes (1×1×2 mm) within the vessel lumen. Thus unless multiple samples are acquired sequentially (a very tedious task) a great deal of information about the flow within the vessel lumen, especially in the region of the carotid bifurcation, will be missed. By modifying the Doppler transducer, replacing the single gate receiver with a linear array multigate, simultaneous velocity measurements can be made across the entire diameter of the vessel. The velocity data is then colour coded and a real-time two-dimensional colour flow image is displayed superimposed on the B-mode image.

The ‘Triplex’ colour coding system assigns red and blue colours to flow velocity vectors of opposite direction (conventionally, the technician usually assigns red to flow in the arterial direction and blue to flow in the venous direction). Variations in Doppler frequency shift are indicated by changes in colour intensity: lighter shades of each colour represent higher flow velocities and maximum velocities appear white.

Triplex examination can provide definition of both normal and abnormal flow patterns in the region of the carotid bifurcation. The usual laminar flow pattern is disrupted at the normal carotid bifurcation, where a zone of boundary layer separation exists along the outer wall opposite the flow divider. In this region normal flow is in the reverse direction and appears as an area of blue on the Triplex scan (Fig. 11) 223. In mild or moderate atherosclerotic disease the plaque will not produce haemodynamically significant stenosis but it will eliminate the normal contour of the bulb, eliminating the normal zone of boundary layer separation. Mild to moderate lesions will cause loss of this area of blue colour at the bifurcation. Haemodynamically significant stenosis will produce stream flow changes with high velocity (white) in the jet and turbulent flow (mixture of colours) just distal to the stenosis (Fig. 12) 224.

The direct studies discussed above provide excellent identification and quantification of extracranial carotid disease, but they do not provide direct information about the actual physiological effects of the carotid stenosis on cerebral perfusion because they fail to measure collateral cerebral blood flow. This is a critical issue, since collateral flow may provide adequate ipsilateral hemispheric perfusion in the presence of a ‘critical’ or haemodynamically significant carotid stenosis, while such a stenosis will have profound detrimental effects in the absence of adequate collateral flow. Direct studies are often supplemented with ocular pneumoplethysmography or supraorbital Doppler studies in an attempt to define collateral flow.

Transcranial Doppler was developed as a method for investigating intracerebral flow directly. A 2-MHz Doppler signal is used to insonate the vessels in the circle of Willis (Fig. 13) 225 and its branches. Four different ‘acoustic windows’—transtemporal, transorbital, suboccipital, and submandibular—are available for insonation of the intracranial vessels (Fig. 14) 226. The transtemporal approach provides access to the anterior cerebral artery including the anterior communicating artery, as well as the posterior cerebral artery and the posterior communicating artery.

Changes in direction and velocity of flow in the various intracranial vessels can verify the presence of significant extracranial disease and can identify the presence of collateral flow. Decreased velocity of flow in the ipsilateral middle cerebral artery confirms the critical nature of the carotid stenosis. Increased velocity and reversal of flow in the contralateral anterior cerebral artery indicates the presence of collateral flow from the contralateral carotid via a patent anterior communicating artery.

The importance of transcranial Doppler as a clinical tool will undoubtedly increase as more experience is gained with the technique and as correlations between haemodynamic findings and clinical outcome are verified.

Non-invasive testing of carotid function provides a safe, convenient, and reliable method for screening patients at risk for stroke, enabling identification of those who might benefit from therapeutic intervention. Patients at risk for stroke include those with a history of transient ischaemic attacks, amaurosis fugax, vertebrobasilar insufficiency, prior stroke, and asymptomatic carotid bruits. Transient ischaemic attacks affecting the carotid distribution may present as hemiparaesthaesia, hemiparesis, and/or speech difficulty. Amaurosis fugax presents as transient loss of vision in one eye. The patient often describes this transient monocular blindness as ‘a shade being pulled down over the eye’. Vertebrobasilar insufficiency is usually characterized by dizziness, bilateral eye symptoms, ataxia, facial numbness, or bilateral extremity weakness or paraesthesiae.

Such symptoms are due to cerebral ischaemia, which may have a variety of possible aetiologies. These include cardiac arrythmias, cardiac thromboembolism, intracranial atherosclerotic disease, and carotid bifurcation disease. There is approximately a 50 per cent chance of the symptoms being due to carotid bifurcation disease, and patients suffering from transient ischaemic attacks have a 5.3 to 8.6 per cent risk of suffering a stroke per year; this is reduced to less than 1 per cent per year following uncomplicated carotid endarterectomy. Symptomatic patients should therefore undergo non-invasive carotid testing; identification of a significant lesion, on the appropriate side for the symptoms should prompt consideration of carotid endarterectomy.

Patients with asymptomatic carotid bruits may also be at increased risk for stroke: some studies have suggested that carotid stenoses of greater than 80 per cent carry a 12 per cent risk of stroke. If the results of non-invasive tests suggest the presence of such a lesion, consideration might be given to prophylactic carotid endarterectomy, although this is a controversial issue.

Until recently, non-invasive studies were used to identify patients with significant extracranial carotid disease; carotid arteriography would then be used to assess the patient for carotid surgery. The non-invasive diagnostic technology has advanced to the stage where redefinition of the indications for preoperative angiography is probably appropriate. In selected patients, such as those with lateralizing hemispheric symptoms and severe haemodynamically significant carotid disease, identified by non-invasive testing, it may be appropriate to proceed to carotid endartectomy without angiographic studies.

A wide spectrum of clinical presentation is seen in patients with peripheral vascular occlusive disease, who may present with intermittent claudication, which may be mild or debilitating, ischaemic rest pain, ischaemic tissue loss, or gangrene. The diagnosis of peripheral vascular occlusive disease as the cause of the symptoms in these patients is often confounded by the presence of associated diseases such as neurological disorders, spinal stenosis (pseudoclaudication) and orthopaedic or rheumatological problems, and diabetes with its associated peripheral neuropathy and neuropathic, non-ischaemic pressure ulcers. Even when an obvious vascular aetiology exists, it is often difficult to determine the exact location of the most significant vascular lesion in the arterial tree. The mere radiological demonstration of a vascular lesion does not confirm its importance: it is the haemodynamic effect on the circulation that is important and this must be confirmed prior to considering any therapeutic intervention. A number of non-invasive studies allow analysis of the haemodynamic significance of any existing vascular pathology. These studies are often helpful in predicting the potential for healing of foot lesions or amputation sites and allow ischaemic and neuropathic foot pain to be differentiated. Haemodynamic studies also provide useful information when the results of peripheral vascular reconstructions or percutaneous angioplasty procedures are being evaluated, and in the detection of failing grafts. The tests also provide a means of assessing and following the progression of any vascular disease which does not require surgical intervention at the time of the patient's initial presentation.

Despite the large variety of studies available three basic parameters are analysed in the study of peripheral vascular disease: pressure, volume change, and flow velocity.

In the normal arterial system the large and distributing arteries produce relatively little resistance to blood flow and there is, therefore, almost no pressure gradient between the aorta and the small arteries of the foot. In fact, as the pressure pulse is propagated distally the pressure pulse wave amplitude and the systolic pressure actually increase due to the decreasing compliance of the distal arterial vessels and the presence of reflected waves. In patients with atherosclerotic peripheral vascular occlusive disease the arterial luminal narrowing which occurs at the sites of atherosclerotic plaque formation produces an increased resistance to flow and a resultant pressure drop across the stenoses. The measurement of distal lower extremity blood pressure therefore provides a method for detecting haemodynamically significant arterial occlusive disease. Since a critical stenosis will produce a pressure gradient across the lesion even before there is any reduction in total flow to the lower extremity, pressure measurement provides a very sensitive method for assessing the presence of physiologically significant arterial disease.

Measuring the systolic arterial pressure at the ankle is a simple means by which any haemodynamically significant arterial lesion from the level of the aorta to the tibial arteries can be identified. The technique involves the placement of a blood pressure cuff around the calf at the level of the ankle. Since Korotkoff sounds are not audible with a stethoscope at this site the pulse is identified with either a continuous-wave Doppler sensor applied over the dorsalis pedis or posterior tibial artery or with a strain gauge plethysmograph or photoplethysmograph applied to the toe. The cuff is inflated well above the systolic arterial pressure and then slowly deflated: the ankle systolic pressure is recorded when the Doppler signal or pulsatile plethysmographic tracing is first identified. The normal ankle systolic pressure is 10 to 20 mmHg greater than the simultaneously recorded brachial systolic blood pressure. Calculating the ratio of the ankle pressure to the brachial pressure provides a convenient index of arterial disease.

The normal ankle:brachial index is greater than 1.0; a value below 0.92 indicates the presence of haemodynamically significant arterial disease. The degree of abnormality of the ankle:brachial index correlates with the location of the occlusive disease, the severity of the stenosis or occlusion, and the number of lesions in the arterial tree. In general, more proximal lesions have a greater effect on the index while higher values are associated with more distal (infrapopliteal) arterial disease. Multilevel disease, the so-called tandem lesion, is associated with a lower ankle:brachial index (Fig. 15) 227. The severity of the patient's symptoms correlates well with the ankle:brachial index, which decreases as the physiological effect of the arterial disease increases. Patients with claudication have a higher ankle:brachial index (0.6–0.8) than those with rest pain (0.26–0.35), while those with ischaemic ulcers have an ankle:brachial index between 0 and 0.25.

Determination of the ankle:brachial index is a simple and reproducible test and is therefore a valuable diagnostic and prognostic tool which may be used to monitor the natural progression of arterial disease in a patient not yet requiring therapeutic intervention. It is also useful in assessing the results of peripheral bypass surgery. Several potential sources of error with this study must be recognized. Repeated determinations of ankle systolic pressures, made on the same day in the same patient, may vary by as much as 14 mmHg. Therefore a change in the ankle:brachial index of more than 0.15 is required before it is safe to conclude that a haemodynamically significant change has occurred. While an abnormal ankle:brachial index indicates the existence of arterial disease, it does not accurately locate the level of the lesion. In a significant proportion of patients with peripheral vascular occlusive disease, and especially in diabetic patients, medial calcification within the wall of the arteries (Monckeberg's sclerosis) produces incompressible vessels. In this instance the cuff pressure required to occlude flow in the underlying artery is greater than the blood pressure because additional pressure is required to collapse the calcified arterial wall. Falsely elevated ankle pressures and ankle:brachial indices are obtained, which underestimate the severity of the arterial occlusive disease. The ankle: brachial index may also underestimate the severity of lesions which produce no pressure gradient at rest but which produce a drop in pressure with the increased flow rates associated with exercise. The addition of other methods of measurement of pressure in the lower extremity increases the diagnostic accuracy of the study and provides further information concerning the severity and location of the arterial disease.

The sensitivity of the ankle:brachial index in diagnosing mild to moderate stenotic disease may be increased by measuring the ankle pressure after exercising the patient. The principle involved is the physiological correlate of Ohm's Law: as the flow across a stenotic lesion increases, the pressure drop across that lesion also increases. Exercise, for example, on a treadmill, causes the peripheral resistance vessels distal to the stenotic lesion to dilate, resulting in an increased flow to the lower limb across the stenosis. The increased pressure drop that occurs with the increased flow can mean that a subcritical lesion at rest becomes haemodynamically significant on exercise.

This test is performed by measuring the ankle:brachial index at rest and again after the patient has exercised on a treadmill for 5 min (or until the patient is forced to stop due to claudication). The patient then resumes the supine position and the quality of the femoral, popliteal, and pedal pulses are immediately reassessed and ankle and brachial pressures are remeasured. Patients with arterial disease will always show a drop in the ankle pressure after exercise, the degree of which correlates with the functional significance of the disease. Multilevel arterial diseases will cause large pressure drops and proximal lesions have a greater impact than distal lesions (an iliac stenosis will produce a greater fall in the post-exercise ankle:brachial index than will a distal superficial femoral artery lesion). This occurs because of the larger muscle mass supplied by the more proximal vessel and the resultant greater diversion of blood flow.

‘Reactive hyperaemia’ may be used as an alternative to exercise to increase the rate of flow across a stenotic segment. This technique involves placing a pneumatic cuff on the thigh and inflating the cuff above systolic pressure for between 3 and 7 min. The temporary ischaemia produces distal vasodilatation; releasing the cuff causes a period of hyperaemic flow. The decrease in pressure that occurs with reactive hyperaemic flow across a stenosis correlates well with that seen following exercise. However, exercise testing is usually the preferred technique since it allows actual observation of the level of activity required to induce claudication and provides the vascular surgeon with a more objective measure of the patient's disability.

Determination of segmental pressure provides greater definition of the atherosclerotic disease process by locating the level at which haemodynamically significant lesions exist. Segmental lower extremity pressures are measured most accurately by applying pneumatic cuffs around the high thigh, low thigh (just above the knee), around the calf just below the knee, and at the ankle. The segmental systolic pressures are then measured in the same manner as that described above.

The normal upper thigh pressure is 30 to 40 mmHg higher than the brachial pressure: this is actually an artefact of the size discrepancy between the thigh and the cuff. The pressure in the cuff is not completely transmitted to the underlying arteries and thus the measured systolic pressure is overestimated. If this apparent elevation in the thigh pressure is not detected then a haemodynamically significant iliac artery lesion or a proximal superficial femoral artery stenosis combined with a profunda stenosis probably exists.

Pressure gradients between any two levels do not normally exceed 20 mmHg, and gradients greater than 30 mmHg between any two segments indicate arterial disease in the intervening arterial segment. If a patient's low thigh pressure is 150 mmHg and the below-knee pressure measurement is 80 mmHg, for example, a popliteal arterial occlusion is probably present.

The technique of indirect non-invasive pressure measurement, by necessity, requires a blood pressure cuff placed at the level being studied and a pulse sensing device (Doppler flowmeter or photoplethysmograph) placed distally. This arrangement creates potential sources of error in interpreting segmental pressure data since segments of arterial disease may lie between the cuff and the sensing device. A proximal superficial femoral artery occlusion, for example, may falsely lower the high thigh pressure measurement, leading to a misdiagnosis of iliac arterial disease. If the cuff thigh size mismatch is excessive (as occurs in patients with large thighs) the apparently elevated high thigh pressure measurement may result in failure to diagnose a significant iliac artery lesion. Accurate assessment of patients with multisegment disease may also be difficult.

A salient component of the preoperative evaluation of the patient with peripheral arterial occlusive disease is the identification of the relative importance of proximal (aortoiliac) or so-called ‘inflow’ and distal (femoropopliteal or tibial) or so-called ‘outflow’ arterial stenoses. A critical assessment of the haemodynamic significance of lesions at the various levels is essential to enable the appropriate level for vascular bypass to be decided. The various lower extremity pressure measurements are very useful in this critical assessment. However, the potential artefacts inherent in the pressure measurements may lead to incorrect conclusions regarding the relative importance of lesions at various levels. Non-invasive arterial flow studies have therefore been developed which may be used along with the pressure measurement data in both diagnosing the presence of arterial occlusive disease and determining the relative importance of lesions present at various levels in the arterial tree. Despite all the advances in lower extremity flow and pressure studies accurate assessment of iliac level disease is often elusive. A variety of indices such as the pulsatility index, damping factors, and inverse damping factors have been developed in an effort to bring better definition to the study of iliac disease. However, they add little to the information contained from the analysis of segmental pressures and flow waveform, discussed below.

Two parameters are available for non-invasive measurement of ‘arterial flow’: these are segmental volume change (plethysmography) and flow velocity (Doppler velocimetry). The data from studies of either of these parameters may be recorded and presented in waveform fashion. Analysis of the waveform ‘pulse contours’ provides valuable information which can be used in a collaborative fashion with segmental pressure studies critically to assess lower extremity arterial disease (Fig. 16) 228.

Plethysmography is the measurement of volume change. A variety of plethysmographs (e.g. mercury strain gauge, impedance and photoplethysmographs) are available, but the most convenient and commonly used device is an air-filled plethysmograph known as the pulse volume recorder. During cardiac systole the part of the limb being examined expands due to arterial inflow; during diastole the venous outflow exceeds arterial inflow and the limb contracts. The pulse volume recorder uses a pneumatic cuff as the segmental volume sensor and measures these changes in volume. Appropriately sized cuffs are placed on the proximal thigh, calf, and ankle (transmetatarsal and digital cuffs may also be used) (Fig. 17) 229. The cuff at the site to be tested is inflated with air to a low pressure of 65 mmHg, bringing the volume sensing cuff into close approximation with the leg. The volume changes in the segment of the leg beneath the cuff produce changes in the air pressure within the cuff bladder; these are then converted to an analogue recording using a pressure transducer. The wave contour obtained closely resembles the waveform of the arterial pulse pressure (Fig. 18) 230.

The contour obtained from pulse volume recording in a normal limb contains four important components: a steep slope in the anacrotic limb, a sharp systolic peak, bowing of the catacrotic limb towards the baseline, and a dicrotic notch. The last represents flow reversal and is due to the normal resistance present in the distal peripheral vascular bed. The presence of a dicrotic notch essentially rules out any haemodynamically significant arterial disease proximal to the pulse volume recorder cuff.

Contour changes that indicate the existence of arterial disease proximal to the cuff include loss of the dicrotic notch, decrease in the rate of rise in the anacrotic limb, rounding and delay in the systolic peak, bowing of the catacrotic limb away from the baseline, and decreased rate of fall in the catacrotic limb. The extent of the pulse wave contour abnormality correlates with the severity of the proximal arterial occlusive disease.

Quantitative as well as qualitative analysis of the pulse volume recorder waveform provides valuable information regarding the severity of the arterial disease proximal to the cuff. Waveform amplitudes are affected by a variety of factors other than arterial occlusive disease, including cardiac output, blood pressure, vasomotor tone, blood volume, muscularity of the limb, and position of the limb. Despite these other variables, progressively worse occlusive disease usually correlates with smaller pulse amplitudes. Five categories have been defined which classify the severity of the occlusive disease based on quantitative and qualitative analysis of the pulse volume recorder waveform (Table 4) 175.

Analysis of the entire profile of segmental pulse volume recorder tracings from the thigh, calf, and ankle levels provides important information regarding the location and severity of occlusive disease. The different patterns resulting when haemodynamically significant disease occurs at different levels are useful in determining the location of physiologically significant disease (Fig. 19) 231.

In the normal lower extremity the amplitude of the calf pulse wave seen on segmental pulse volume recording is larger than the thigh pulse wave. The phenomenon, which has been termed ‘augmentation’, is actually an artefact which results from the smaller size of the calf cuff compared to the thigh cuff: volume change produces a relatively greater pressure change in the smaller cuff than in the larger cuff (Fig. 20) 232. The absence of the calf augmentation phenomenon indicates the presence of superficial femoral artery occlusive disease (Figs. 21, 22) 233,234.

In the presence of isolated iliac artery disease with an open distal system the pulse contours at all the levels on the affected side will be abnormal, but calf augmentation will still exist (Fig. 23) 235. In the presence of occlusion of the superficial femoral artery and stenosis of the profunda femoral artery, segmental wave contours may lead to a false diagnosis of iliac artery disease: analysis of pulse volume recordings made at thigh level after exercise will usually allow differentiation between the two conditions. If an iliac lesion exists the post-exercise thigh pulse contour will deteriorate further, while a normal artery and diseased superficial femoral artery will produce no significant change in the contour after exercise. The degree of accuracy with which severity and location of arterial lesions can be assessed is greatly increased when analysis of pulse volume recordings is combined with segmental pressure studies.

Haemodynamically significant arterial stenosis produces changes in the arterial flow velocity and in the velocity pulse wave contour and disrupts laminar flow, creating turbulence. Each of these effects can be studied and the severity of the lesion estimated by Doppler examination. The same Doppler technology and methods used in the assessment of carotid disease may be applied to the study of lower extremity arterial disease. Unlike pressure measurements and plethysmography, which are indirect studies, Doppler studies provide direct information about the arterial disease.

A ‘pencil’ probe is most commonly used (Fig. 24) 236, and is placed directly over the vessel to be insonated, maintaining a 45 to 60° angle of incidence to the vessel. To examine the common femoral artery the probe is positioned at or above the inguinal ligament to avoid inadvertent insonation of the superficial femoral or profunda femoral artery. The superficial femoral artery is best approached with the probe in a medial position and directed between the quadriceps and adductor muscles. The popliteal artery is located with gentle flexion and external rotation of the leg. The posterior tibial artery is insonated in its position just behind the medial malleolus. The dorsalis pedis and lateral tarsal (terminal branch of the peroneal) arteries are located slightly lateral to the extensor hallucis longus tendon and anteromedial to the lateral malleolus, respectively. Although these are the standard positions examined during the routine Doppler study of the lower extremities, arteries may be insonated at virtually any point along their course in the leg. This provides a versatile tool for directly examining arterial stenoses anywhere in the lower extremity. Where the arteries lie in a superficial location, a 10-MHz probe is used; arteries in deeper positions (e.g. peroneal) must be examined with a 5-MHz probe.

The audible Doppler signal provides important information which may suggest the presence of a haemodynamically significant stenosis. However, this type of Doppler signal analysis is subjective, and it provides no hard copy of the results. Most continuous-wave Doppler instruments used for peripheral arterial studies are equipped with zero-crossing frequency to voltage converters which produce a velocity analogue waveform which is recorded on a strip chart. The contour of the analogue waveform closely resembles that of the arterial pressure pulse (Fig. 25) 237. However, this method may produce errors since the voltage output is proportional to the root mean square frequency rather than to the mean Doppler shift frequency, resulting in overestimation of low velocities and underestimation of high velocities. Despite these limitations, the analogue waveform provides excellent qualitative information about the arterial flow.

The normal analogue waveform is triphasic, with a rapid increase in velocity in early systole, followed by a rapid fall in velocity with reversal of flow in early diastole. In late diastole, there is a smaller forward flow component (Fig. 25(a)) 237.

Arterial stenoses proximal to the segment being insonated will produce changes in the contour of the analogue waveform which correlate with the haemodynamic significance of the lesion. The earliest sign of a proximal stenosis is loss of the diastolic flow reversal, resulting in a biphasic waveform. As the severity of the stenosis increases, the slopes of the upstroke and downstroke decrease and the peak becomes rounded rather than peaked. Progressive dampening of the waveform results in a low amplitude monophasic waveform seen in patients with severe stenoses (Fig. 25(b–g)) 237. Analysis of both the segmental analogue waveforms and pressures will usually provide sufficient data to enable haemodynamically significant arterial diseases to be diagnosed and will often allow accurate assessment of the location of the disease (Fig. 26) 238.

Doppler velocimetry measurements can be made with far greater accuracy and precision when combined with duplex scanning technology. This greater definition is achieved at the price of greater technical complexity and longer time required to perform the study. Lower extremity arterial duplex scanning is performed by longitudinal imaging of the femoral, popliteal, and tibial arteries using real-time B-mode ultrasound. At the infrapopliteal level, the technical ease of the study is increased if triplex (colour duplex) capability is available: the colour technology reduces the acquisition time needed to locate the tibial vessels. Doppler flow velocity measurements are then obtained from the iliac, common femoral, superficial femoral, profunda femoral, popliteal, and tibial arteries. If the B-mode image visualizes any stenotic regions the sample volume of the pulsed Doppler may be placed directly into the region of the stenosis. This powerful diagnostic tool allows velocity measurements to be made in the areas most likely to be abnormal rather than simply at random sites where the flow may have already returned to normal. The Doppler frequency shift data are then analysed using fast Fourier transform to produce a spectral analysis which contains more information and is more accurate than the zero-crossed Doppler analogue waveform.

The normal peripheral arterial velocity spectral waveform is triphasic, with a contour similar to that of the analogue waveform. A pulsed Doppler sample volume obtained from centre stream flow yields a spectrum with a narrow frequency bandwidth and a systolic window, represented as a clear area in the waveform beneath the systolic peak. Stenoses produce an increase in peak systolic velocity and spectral broadening due to turbulent flow. Haemodynamically significant lesions (stenosis above 50 per cent) cause loss of diastolic reversed flow and an increase in diastolic forward flow. Analysis of waveform contour, peak systolic velocity, presence of diastolic reverse flow, peak diastolic forward flow, and spectral broadening allows arterial stenoses to be classified into five categories: normal, less than 20 per cent diameter reduction, 20 to 49 per cent diameter reduction, 50 to 99 per cent diameter reduction, and occlusion (Table 5) 176.

A normal waveform is triphasic with a clear systolic window. When a minimal stenosis is present (less than 20 per cent diameter reduction) the waveform contour and peak systolic velocity are unaltered, but mild spectral broadening is seen. Stenoses in the moderate range (20 to 49 per cent) cause a 30 to 50 per cent increase in the peak systolic velocity and more marked spectral broadening, with filling in of the systolic window. The reverse flow component is usually still present. When severe arterial narrowing occurs (diameter reduction of 50 to 99 per cent) the lesion becomes haemodynamically significant. The pressure drop across the lesion results in reduced peripheral resistance, with resultant loss of the reversed flow component, and the peak diastolic forward flow is elevated. The waveform contour becomes monotonic and there is often a more than 100 per cent increase in peak systolic flow velocity (Fig. 27(a–d)) 239.

The studies reviewed in this chapter all provide haemodynamic evidence of the degree of arterial insufficiency that exists in a patient with peripheral vascular occlusive disease. Transcutaneous oximetry can provide valuable metabolic data which may be used to supplement the haemodynamic data provided by the other non-invasive tests. Transcutaneous oxygen tension provides information about the metabolic state of the tissue being studied and a measure of the adequacy of the arterial oxygen supply to the tissues.

Measurement of transcutaneous oxygen tension involves the placement of an oxygen-sensing electrode on the skin. The device used contains a small heating unit which warms the skin to a temperature conducive to efficient oxygen diffusion. The quantity of oxygen available for diffusion to the skin is a function of the arterial flow to the area and the amount of oxygen extracted from the blood to meet the metabolic requirements of the tissues. When arterial occlusive disease is severe the tissue perfusion becomes marginal and capillary oxygen perfusion decreases as the proportion of oxygen extraction must increase to meet the metabolic demands of the tissue.

Transcutaneous oxygen tension is usually measured at the dorsum of the foot, medial aspect of the calf, and the thigh, with a reference electrode placed in the infraclavicular region. Comparing the value obtained at the lower extremity to that at the chest yields an index which can be compared to a normal value with regard to the patient's age, cardiac output, and arterial oxygen tension, as well as other factors. The normal value is approximately 60 mmHg, but this may decrease by 5 to 6 mmHg in a normal lower extremity. The normal transcutaneous oxygen tension index is 0.9.

Measurements of transcutaneous oxygen tension are not affected by mild or moderate arterial insufficiency; however severe peripheral vascular occlusive disease will produce significant changes in the oxygen tension measured at the foot. The study is therefore, often valuable in assessing a patient's risk for limb or tissue loss or in differentiating ischaemic rest pain from neuropathic pain: values of less than 20 mmHg, measured at the foot, are usually obtained in patients with rest pain, ischaemic ulcers, or gangrene.

Exertion-related lower extremity pain may be due to arterial disease, neurological disorders (pseudoclaudication), and orthopaedic or rheumatological problems. Segmental pressures and pulse volume recordings or Doppler velocity waveforms will differentiate vascular from other aetiologies. If ankle pressures greater than 50 mmHg and pulse volume recording categories of 2 or 3 (Table 4) 175 are found, vascular claudication is unlikely. If the ankle pressure is less than 50 mmHg and the pulse volume recording category is 4 or 5, a vascular aetiology is likely. Exercise studies are very important in determining the cause of exertional leg pain: changes in the distal pressure and pulse volume or Doppler waveforms often confirm the diagnosis (Fig. 24) 236.

Less than 5 to 8 per cent of claudicants will progress to the situation where limb loss is inevitable. However, severe claudication can be incapacitating, and the non-invasive studies are often used to supplement the history and physical examination in deciding whether surgical intervention is required.

In patients who present with foot pain at rest, the primary function of the vascular study is to differentiate neuropathic from ischaemic foot pain: this is often difficult in diabetic patients. The pertinent studies include measurement of ankle and toe pressures and analysis of pulse volume recordings or Doppler waveforms, as well as measurement of transcutaneous oxygen tension. Ankle pressures less than 35 mmHg (less than 55 mmHg in diabetics) and pulse volume recording categories of 4 or 5 essentially confirm the diagnosis of ischaemic rest pain. If the ankle pressure is greater than 55 mmHg (greater than 80 mmHg in diabetic patients) and pulse volume recordings are in categories 1 to 3, ischaemic pain is unlikely. In diabetics with non-compressible vessels, the pulse volume or Doppler waveforms are more useful than the actual pressures, which may be deceptively high. Measurement of transcutaneous oxygen tension is also valuable in the assessment of the diabetic since non-compressibility of vessels has no effect on this measurement. If ischaemic pain is confirmed, the patient is at high risk for ischaemic tissue (limb) loss and should be evaluated for vascular bypass procedure.

Foot ulcers may be due to either ischaemic or neuropathic disease. When an ischaemic aetiology is confirmed by non-invasive studies, the results can also predict the probability of successful healing of the lesion. An ulcer in a patient with an ankle pressure of 30 to 40 mmHg is not likely to heal unless a vascular bypass can improve the perfusion pressure in the foot. Toe pressure measurements are more accurate than ankle pressures in predicting the probability of successful ulcer healing: pressures greater than 30 mmHg predict a high likelihood of successful healing with conservative management. Toe pressures less than this predict the need for revascularization to effect successful healing of the lesion.

Distal bypass grafts may fail for a variety of reasons, including retained venous valves, arteriovenous fistulae, anastomotic intimal hyperplasia, and progression of atherosclerosis in the graft or native arteries. Technical errors, myointimal hyperplasia, and progression of atherosclerotic disease cause graft failure by reducing graft blood flow below the thrombotic threshold velocity, below which thrombosis of the graft occurs. Postoperative surveillance allows the failing graft to be identified before occlusion occurs. Indirect physiological measurements such as ankle: brachial index segmental pulse volume recording or Doppler velocity analogue waveforms are probably not sensitive enough to detect reliably the failing but patent graft. Duplex scanning provides a very critical tool for such graft surveillance, providing both B-mode imaging, which defines anatomic complications of the graft (i.e. retained valves or intimal hyperplasia), and Doppler velocimetric data, which may identify a low flow state before thrombosis actually occurs. A low flow state exists when the peak systolic flow velocity is less than 45 cm/s or when a drop in velocity of greater than 30 cm/s occurs between serial measurements in the same graft. An abnormal B-mode scan or velocity measurement should prompt investigation with an arteriogram to assess the graft for possible revision.

Aaslid R, Markwalder TM, Nornes H. Noninvasive transcranial doppler ultrasound recording of flow velocity in basal cerebral arteries. J Neurosurg 1982; 57: 769.

Baker JD, Dix D. Variability of doppler ankle pressures with arterial occlusive disease: an elevation of ankle index and brachial-ankle gradient. Surgery 1983; 89: 134.

Baker JD, Barker WF, Machledeer HI. Evaluation of extracranial cerebrovascular disease with ocular pneumoplethysmography. Am J Surg 1978; 136: 206–8.

Baker WH. et al. Diagnosis of peripheral occlusive disease: comparison of clinical evaluation and noninvasive vascular laboratory. Arch Surg. 1989; 11: 1308–10.

Beach KW, Phillips DJ. Doppler instrumentation for the evaluation of arterial and venous disease. In Jaffe CC, ed. Vascular and Doppler Ultrasound. New York: Churchill Livingstone, 1984: 11–49.

Berkowitz HD, et al. Value of routine vascular laboratory studies to identify vein graft stenosis. Surgery 1981; 90: 971–9.

Bernstein NM, Beloev ZG, Norris JW. The limitations of diagnosis of carotid occlusion by doppler ultrasound. Ann Surg 1988; 207: 315–7.

Blackshear WM, et al. Detection of carotid occlusive disease by ultrasound imaging and pulsed doppler spectrum analysis. Surgery 1979; 86: 698–706.

Brandyk DF. Monitoring after distal arterial reconstruction. In Bernstein EF, ed. Recent Advances in Noninvasive Diagnostic Techniques in Vascular Disease. St Louis: CV Mosby 1990: 247–58

Call GK, et al. Correlation of continuous-wave doppler spectral flow analysis with gross pathology in carotid stenosis. Stroke 1988; 19: 584–8.

Carter SA. Role of pressure measurements in vascular disease. In Bernstein EF, ed. Noninvasive Diagnostic techniques in Vascular Disease. 3rd edn. St Louis: CV Mosby, 1985: 513–44.

Eikelboom BC. Ocular pneumoplethysmography. In: Bernstein EF, ed. Noninvasive Diagnostic Techniques in Vascular Disease. 3rd edn. St Louis: CV Mosby 1985.

Gee W, Mehigan JT, Wylie EJ. Measurement of collateral hemispheric blood pressure by ocular pneumoplethysmography. Am J Surg 1975; 130: 121–7.

Goodson WF, Flanigan DP, Bishara Ra, Schuler JF, Kikta MJ, Meyer JP. Can carotid duplex scanning supplant arteriography in patients with facial carotid territory symptoms? J Vasc Surg 1987; 5: 551–7.

Hausser CJ, Shoemaker WC. Use of a transcutaneous Po&sub2; regional perfusion index to quantify tissue perfusion in peripheral vascular disease. Ann Surg 1983; 197: 337–43.

Harward TR, Bernstein EF, Fronek A. Continuous-wave versus range-gated pulsed doppler power frequency spectrum analysis in the detection of carotid arterial occlusive disease. Ann Surg 1986; 204: 32–7.

Kameyaama M, Okinaka S. Collateral circulation of the brain with special reverence of atherosclerosis of the major cervical and cerebral arteries. Neurology 1963; 13: 270–86.

Kempczinski RF. Segmental volume plethysmography in the diagnosis of lower extremity arterial occlusive disease. J Cardiovasc Surg 1982; 23: 125–9.

Kempczinski RF. Segmental volume plethysmography: the pulse volume recorder. In Kempczinski RF, Yao JST, eds. Practical Noninvasive Vascular Diagnosis, 2nd edn. Chicago: Year Book Medical Publishers, 1987: 142–53.

Kohler TR et al. Duplex scanning for aortoiliac and femoropopliteal disease: a prospective study. Circulation 1987; 76: 1074–80.

Leopold PW. et al. Duplex ultrasound: its role in the noninvasive follow-up of the in situ saphenous vein bypass. J Vasc Technol 1987; 11: 183.

Megerman J, Abbott WA. Transcutaneous oxygen tension determination. In Kempczinski RF, Yao JST, eds. Practical Noninvasive Diagnosis. 2nd edn. Chicago: Year Book Medical Publishers 1987: 210–29.

Minh-Chau L, et al. Decreased graft flow velocity is a reliable predictor of impending failure of reversed vein grafts. J Vasc Technol 1988; 12: 133.

Moll FL, Eikelboom BC, Vermeulen FE, VanLier JJ, Schulte BP. Dynamics of collateral circulation in progressive asymptomatic carotid disease. J Vasc Surg 1986; 3: 470–4.

Moneta GL, Taylor DC, Strandness E. Noninvasive assessment of cerebrovascular disease. Ann Vasc Surg 1986; 1: 489–501.

Moneta GL, et al. Operative versus nonoperative management of asymptomatic high grade internal carotid artery stenosis: improved results with endarterectomy. Stroke 1987; 18: 1005.

Neumyer MM, Thiele BL. Evaluation of lower extremity occlusive disease with doppler ultrasound. In Taylor KJW, Burns PN, Well PNT, eds. Clinical Applications of Doppler Ultrasound. New York: Raven Press. 1988; 317–37.

Oh PIT, Provan JL, Amelie FM. The predictability of the success of arterial reconstruction by means of transcutaneous oxygen tension measurements. J Vasc Surg 1987; 5: 356–62.

Otis SM, Ringelstein EB. Principles and applications of transcranial doppler sonography. In Bernstein EF, ed. Recent Advances in Noninvasive Diagnostic Techniques in Vascular Disease. St Louis: CV Mosby, 1990: 59–79.

Ramsey DE, Manke DA, Sumner DS. Toe blood pressure—a valuable adjunct to ankle pressure measurement for assessing peripheral arterial disease. J Cardiovasc Surg 1983; 24: 43–8.

Spencer M. Continuous-wave doppler imaging of the carotid bifurcation. In Bernstein EF, ed. Noninvasive Diagnostic Techniques in Vascular Disease. 3rd edn. St Louis: CV Mosby 1985: 367–9.

Strandness DE, Jr. The Seattle data. In Bernstein EF, ed. Recent Advances in Noninvasive Diagnostic Techniques in Vascular Disease. St. Louis: CV Mosby 1990: 129–34.

Sumner DS. Noninvasive assessment of peripheral arterial occlusive disease. In Rutherford RB, ed. Vascular Surgery. Eastbourne: WC Saunders 1989; 61–111.

Sumner DS. Noninvasive assessment of peripheral arterial occlusive disease. In Rutherford RB, ed. Vascular Surgery. Eastbourne: WB Saunders 1989: 61–3.

Sumner DW, Strandness DE, Jr. Hemodynamic studies before and after extended bypass grafts to the tibial and peroneal arteries. Surgery 1978; 84: 348.

Turnipseed WD, Acker CW. Postoperative surveillance. An effective means of detecting correctable lesions that threaten graft patency. Arch Surg 1985; 120: 324–8.

Wells PNT. Instrumentation including color flow mapping. In Taylor KJ, Burns PN, Wells PNT, eds. Clinical Applications of Doppler Ultrasound. Raven Press 1987: 26–45.

Wieber DO, Whisnant JP. In Warlow C, Morris PJ, eds. Transient Ischemic Attacks. New York: Marcel Dekker, 1982; Ch. 1: 1–19.

Yao JST. Noninvasive techniques of measuring lower limb arterial pressures. In Bernstein EF, ed. Noninvasive Diagnostic Techniques in Vascular Disease. 3rd edn. St Louis: CV Mosby 1985: 83–90.

Zierler RE. Color-flow doppler imaging: normal and abnormal carotid studies. In Bernstein EF, ed. Recent Advances in Noninvasive Diagnostic Techniques in Vascular Disease. St Louis: CV Mosby, 1985: 37–41.

Zierler RE, Strandness DE, Jr. Duplex scanning for diagnosis of aortoiliac and lower extremity disease. J Vasc Technol 1987; 11: 99–102.