The use of venous occlusion plethysmography to measure blood flow in humans was first described over 90 years ago by Hewlett & van Zwaluwenburg . Although initially limited by the need for cumbersome equipment, the advent of mercury-in-silastic strain gauges over 50 years ago greatly increased its applicability, and the technique has remained essentially unchanged since except for the introduction of computerization. Venous occlusion plethysmography has been used extensively to study human vascular physiology in vivo, and is at its most powerful when combined with intra-arterial drug administration, usually into the forearm vascular bed. This permits detailed study of both vascular pharmacology, usually by infusion of receptor agonists, and physiology, by infusion of receptor antagonists. Indeed, forearm venous occlusion plethysmography with local brachial artery infusion has become one of the ‘gold-standards’ in the assessment of vascular function in health and disease, and an accurate, reproducible and convenient method with which to assess the effect of new vasoactive drugs and hormones in humans in vivo. Besides being minimally invasive, the major advantage of this approach is that drugs can be administered at subsystemic doses minimizing disturbances to systemic physiology. Although venous occlusion plethysmography can be equally well applied to the lower limb, this review will focus on the forearm vascular bed.
Measurement of blood flow
The underlying principle of forearm venous occlusion plethysmography is simple: when venous drainage from the arm is briefly interrupted, arterial inflow is unaltered and blood can enter the forearm but cannot escape. This results in a linear increase in forearm volume over time, which is proportional to arterial blood inflow, until venous pressure rises towards the occluding pressure . Under resting conditions, ∼70% of total forearm blood flow (FBF) is through skeletal muscle, with skin blood flow accounting for most of the remainder . However, the hand contains a high proportion of arterio-venous shunts and, because skin blood flow is highly dependent on temperature  and has a different basic pharmacology and physiology than muscle blood flow, it is standard practice to exclude the hands from the circulation during measurement of forearm blood flow. Indeed, if the hand is not excluded then blood flow is nonlinear .
The basic methodology has changed little since its first description by Hewlett & Zwaluwenburg over 90 years ago . Venous return from the forearm is briefly interrupted by inflating a cuff, placed around the upper arm, to well above venous pressure but below diastolic pressure. Typically, an inflation pressure of around 40 mmHg is used for intervals of 10 s, followed by 5 s of deflation, which does not alter arterial inflow and allows venous emptying . The forearm must be positioned above the level of the heart to ensure adequate venous emptying during the period of deflation, which is usually achieved by resting the elbows on foam pads and supporting the hands with pillows (Figure 1). The hands are excluded from the circulation during measurements by initial rapid inflation of a smaller cuff, placed around the wrist, to well above systolic pressure (220 mmHg for normotensive subjects). The wrist-cuffs must be inflated at least 60 s before starting measurements of flow in order to allow FBF to stabilize . As this manoeuvre renders the hand ischaemic, the measurement period is limited, but periods of up to 13 min have been employed safely . Automated venous occlusion plethysmography equipment has been developed  and is of particular use when on-line measurement of flow is required, as discussed later.
Changes in forearm volume are measured by a plethysmograph. Initially, air and then water-filled jackets were used, but these have been largely replaced by mercury-in-rubber (or silastic) strain gauges , which may themselves be ultimately replaced by indium-gallium gauges due to concerns over the potential toxicity of mercury. The strain gauges should be placed around the widest part of the forearm, and simply act as resistors connected as one arm of a Wheatstone bridge . Changes in forearm volume result in a corresponding change in arm circumference and thus strain gauge length, which can be detected as an alteration in electrical resistance of the gauge, and thus potential difference (Figure 2) . If the gauge length is made equal to the resting circumference of the limb, then changes in limb volume are directly proportional to the changes in resistance .
Venous occlusion plethysmography provides a measure of blood flow to that part of the forearm enclosed by the two cuffs. This is usually expressed as ml per 100 ml of forearm volume per minute, when electronic calibration is employed [6, 12]. Actual forearm volume can be estimated by calculation, assuming the forearm is a simple truncated cone, or by simple water displacement. When studies are conducted in a quiet, temperature-controlled room (22–26 ° C), with the subject resting in a comfortable supine position, measurement of FBF by strain gauge venous occlusion plethysmography compares favourably with values obtained by other established techniques [13, 14]. More recently, strain gauges employing a nonelastic loop attached to a passive inductive transducer, which acts as an electromagnetic sensor, have been developed and appear to be at least as accurate as mercury gauges .
Alternatives to plethysmography
Blood flow can also be measured using ultrasound. Usually this involves combining estimates of mean blood velocity with the cross sectional area of the vessel. Doppler ultrasound is used to estimate the mean blood velocity, and B-mode ultrasound used to estimate the vessel diameter. Measurement of mean velocity is most accurate when the ultrasound beam-width is greater than the vessel diameter; with modern ultrasound scanners the 3dB beam width is typically 1–2 mm, so that this condition is not met in most vessels of interest, leading to an overestimation of mean velocity. Assessment of vessel diameter relies on the ability to image the anterior and posterior vessel walls and depends on the angle of the probe to the artery. As the area is obtained by squaring the diameter, any error in diameter estimation results in doubling of the error for area; in general it is difficult to measure area in vessels less than about 3–4 mm in diameter, though this is dependent on the location of the vessel and the frequency of the transducer used, with greater accuracy being achieved if high frequency transducers are used for scanning superficial vessels. Typically, in vessels of 5–10 mm in diameter, the flow may be measured at best with an accuracy of ∼20–30%. In routine practice, using the above technique, it is easy to generate very large errors of ∼50 or even 100% in flow estimation, but in the experimental setting improved accuracy, below 10%, may be possible . Overall, the use of ultrasound does not confer any particular advantage over plethysmography in assessing blood flow, although it does allow conduit vessel diameter to be determined at the same time. Moreover, ultrasonic methods require expensive equipment and highly skilled operators and can be difficult to use with intra-arterial needles.
Minimum vascular resistance
Following limb ischaemia there is a rapid increase in forearm blood flow, which slowly returns to baseline values and is termed reactive hyperaemia. The maximum flow and area under the curve for this increase in blood flow are directly related to the duration of ischaemia  until ∼13 min, after which lengthening the period of ischaemia does not increase the maximum recorded FBF further . Indeed, the maximal post-ischaemic FBF is not influenced by sympathetic tone  or administration of vasodilators  and seems to represent the minimum forearm vascular resistance, which is a function of the average wall:lumen ratio of the resistance vessels . Therefore, minimum vascular resistance provides indirect information concerning the structure of resistance vessels in the forearm. Moreover, minimum forearm vascular resistance correlates well with more direct measurements of resistance vessel structure, such as the media:lumen ratio . Increased minimum vascular resistance has been demonstrated in individuals with established hypertension [19, 21], young men with ‘borderline’ hypertension , and those with ‘white-coat’ hypertension . Interestingly, such structural changes can be reversed with antihypertensive medication [23–25].
Minimum vascular resistance can be assessed reproducibly in the forearm following 13 min of ischaemia, induced by inflating a cuff sited around the upper arm to 300 mmHg for 13 min, and asking subjects to perform 20–30 hand contractions during the last 1 min of ischaemia . As is usual, the hands should be excluded from the circulation before starting the study and measurements of FBF, using venous occlusion plethysmography, should be made at 15 s intervals for 3 min following ischaemia. Maximum FBF usually occurs between 5 and 45 s after cuff deflation and, practically, vascular resistance is calculated as mean arterial pressure÷FBF. Although this technique may seem objectionable, it is generally well tolerated.
Forearm venous volume
The compliance of the deep forearm veins can be assessed using venous occlusion plethysmography . Moreover, this can be performed during intra-arterial drug infusions, to provide simultaneous information regarding the effect of drugs on resistance vessels and veins [27–29]. Several techniques exist for determining venous distensibility. However, one of the simple approaches is to increase the upper arm cuff pressure in a stepwise manner to ∼30 mmHg, at intervals of ∼1 min. Under these conditions, changes in forearm volume reflect changes in venous volume, and thus pressure-volume curves can be constructed to determine venous distensibility (Figure 3).
Following inflation of the upper arm cuff to above venous pressure, there is a rapid rise in forearm volume, which reaches a plateau after 1–2 min, depending on the inflation pressure used . However, during prolonged venous occlusion there is a small but measurable continued increase in forearm volume (Figure 4). This appears to be due to extravasation of fluid from the capillaries, rather than progressive venous distension [31, 32] and, therefore, venous occlusion plethysmography can be used to assess capillary permeability . The pressure in the occluding cuff may be raised in a single step , or in progressively by small increments , in which case fluid flux is calculated by least-squares fitting of the volume response after the vascular compliance component is completed . Changes in microvascular permeability have been described in diabetic patients  and in response to infused peptides , as illustrated in Figure 5.
Sympathetic nervous system activity
Application of a small degree of negative pressure (10–15 mmHg) to the lower body produces venous pooling in the legs, and unloading of the baroreceptors . This results in increased sympathetic activity in the nerves supplying the upper limbs and an ∼20% reduction in FBF, without any concomitant change in arterial pressure or heart rate [36, 37]. This technique has been used to examine the influence of drugs on peripheral sympathetic function. For example, angiotensin II, in low doses, does not significantly alter FBF, but markedly enhances the sympathetic-mediated vasoconstriction to lower body negative pressure .
Local drug administration
Perhaps one of the most important uses of venous occlusion plethysmography has been the study of the local effect of vasoactive mediators or drugs in the forearm vascular bed. Drugs can be given intra-arterially by placement of a fine (27-gauge) needle into the brachial artery under local anaesthesia, which causes minimal discomfort. This is a simple, safe technique and over 90% of subjects can be cannulated at the first attempt. Indeed, a recent survey canvassing experienced users from around the world found that several thousand cannulations have been conducted over the last 15 years without any serious adverse events (P. Vallance; personal communication). Nevertheless, the technique of cannulation is best learnt from an experienced researcher, and general precautions concerning sterility and the preparation of drugs for infusion need to be carefully observed. Infrequent minor problems at the time of cannulation, such as bruising and local discomfort, usually settle rapidly without any problems. However, it is our current practice to always contact the subject the day after cannulation to ensure that they have not experienced any delayed problems. Larger needles can be employed, but there seems no particular advantage of this technique, in most situations, and it is generally less well tolerated and does not allow for frequent repeat studies. Indeed, serial studies can be undertaken safely with the use of a 27-gauge needle , and individual investigators (JRC, DJW) have personally undergone > 50 cannulations without harm. The particular advantage of the technique is that drugs can be given at doses 10–100 times below those causing systemic effects because FBF (< 50 ml min−1) is ∼100 fold lower than cardiac output. Therefore, if long-acting drugs are used for brief infusion periods, or short-acting drugs for longer periods, their effects on resistance vessels can be studied without invoking systemic effects or neurohormonal reflexes , minimizing the potential for side-effects and toxicity. Moreover, the non-infused arm can be used as a contemporaneous control throughout the experiment  to take into account any minor changes in blood flow affecting both arms, such as emotional stress or changes in temperature . It is now rare for unilateral studies to be performed, and this technique should not be encouraged.
Several groups have assessed the reproducibility of venous occlusion plethysmography. Roberts et al. demonstrated good within-subject reproducibility for unilateral blood flow measurement, with a coefficient of variation of ∼10% for FBF measurement and ∼11% for calf blood flow under resting conditions, and 13% for calf blood flow post-exercise. Although others have reported somewhat higher values of ∼25% for assessment of unilateral FBF over several weeks . Petrie et al. investigated the within-subject reproducibility of bilateral forearm venous occlusion plethysmography over three visits 1 week apart, and reported a coefficient of variation of 31–39% for unilateral flows at rest, but only 19% when the ratio of unilateral flows was used. During infusion of the vasoconstrictors angiotensin II and noradrenaline reproducibility of the response was again improved by analysing the results as percentage change in FBF ratio (14 vs 18% and 16 vs 27%, for interquartile ranges, respectively). However, Walker et al. found that the intra-subject variability of the response to dilators (acetylcholine and salbutamol) was much less when quoted as absolute values of FBF than when analysed as percentage change in the FBF ratio. The same also appears to be true for exercise-induced increases in FBF .
When only unilateral FBF is measured, results may be expressed in absolute terms, i.e. ml per 100 ml of forearm volume per minute or as a change from baseline. However, when, as is usual practice, blood flow is measured in both arms simultaneously, several approaches to analysing and presenting these data exist. Again, FBF may be expressed in absolute values for each arm, the control arm merely providing evidence concerning any systemic affects, the occurrence of which would invalidate such an approach. Changes may be reported either in absolute terms or as a percentage of baseline values [46–48]. Alternatively, the ratio of FBF in the infused arm to the control arm (usually around 1) may be quoted, frequently as a percentage change from baseline. This approach was first suggested by Greenfield & Patterson , and has been widely used by others since [38, 50, 51]. It has the theoretical advantage of minimizing the influences of small changes in FBF affecting both arms due, for example, to alteration in sympathetic activation or blood pressure ; the non-infused arm acting as a contemporaneous control, and appears to give better reproducibility than reporting unilateral flow, at least for vasoconstrictors . However, with vasodilators, especially those that can produce large increases in blood flow, citing changes in absolute blood flow may be better since small changes in the flow in the control arm can have a large influence on the percentage change in the FBF ratio at high flows, increasing interstudy variability . It should be noted that in studies where several different vasoactive agents are infused sequentially it is important to use the flow immediately preceding each infusion as the baseline to assess changes in FBF rather than the original baseline at the start of the study. Flows may also be used to calculate forearm vascular resistance, and results expressed in this format [53, 54]. However, this approach makes several important assumptions, such as assuming laminar flow and that alterations in blood pressure have no effect on the state of smooth muscle contraction, and is, therefore, we believe inappropriate . Indeed, calculating forearm vascular resistance confers no particular advantage over FBF values and may, on occasion, be misleading, especially if baseline blood pressure differs between study groups.
The most frequently employed method for statistical analysis of FBF data is repeated measures analysis of variance, with posthoc Student's t-tests. Alternatively, summary measures statistics may be used, such as maximal response , which may be of more use when the response is biphasic, or area under the curve. However, the exact choice of method may well depend on the question one wishes to address.
Local drug and hormone infusions
Venous occlusion plethysmography, coupled with brachial artery drug administration, provides an ideal method with which to assess the local effect of drugs and hormones on peripheral resistance vessels without invoking systemic effects (Table 1, Figure 6). Whereas, receptor agonists are usually employed to investigate pharmacological mechanisms, antagonists often provide more useful information regarding physiology of the system under investigation. When given systemically, many vasoactive drugs produce changes in blood pressure and sympathetic tone that will alter FBF [37, 41, 56]. Moreover, the lack of systemic effects allows near full dose response curves to be constructed for some agonists, such as angiotensin II  (Figure 6) and atrial natriuretic peptide . However, for some vasodilators, such as bradykinin (unpublished data) and substance P , construction of a full dose–response curve is not possible due to ‘spill over’ from the arm and consequent systemic effects at high doses. Moreover, large increases in flow will also reduce the effective concentration of a drug in the forearm due to dilution and potentially confound interpretation of the results, although this does not appear to be a problem in practice. However, limited dose–response curves can still be constructed which may be usefully employed to assess the effect of receptor antagonists  or interventions such as inhibition of angiotensin converting enzyme , or endothelin converting enzyme .
Table 1. Vasoactive peptides and drugs in human forearm resistance vessels.
Atrial natriuretic peptide
Calcitonin gene-related peptide
BQ788 (ETB receptor anatagonist) Phentolamine
Acetylcholine and carbachol
Sodium nitroprusside and glyceryl trinitrate
BQ123 (ETA receptor antagonist)
Endothelin converting enzyme inhibitors
Neutral endopeptidase inhibitors
The FBF technique avoids potential risks involved with systemic administration of vasoconstrictors and although critical closure of the forearm vascular bed is a theoretical possibility with potent vasoconstrictors , it has not been reported, even with large reductions in FBF observed with endothelin-1  and angiotensin II , when suitable doses and measurement periods are employed. Moreover, the response of the forearm vascular bed to locally infused drugs correlates well with systemic effects for many agents including angiotensin II [40, 62], endothelin-1 [63, 64] bradykinin [65, 66], and sodium nitroprusside [67, 68]. In contrast, when drugs are given systemically, the observed changes in local blood flow do not always mirror those seen during local intra-arterial infusion. Indeed, despite angiotensin II being a powerful vasoconstrictor when administered directly into the forearm vascular bed, systemic infusion actually increases calf blood flow at high doses even though this is accompanied by a rise in blood pressure and systemic vascular resistance . Therefore, changes in FBF cannot be used to reliably indicate the effects of drugs on resistance vessels when such agents are infused systemically.
Measurement of FBF coupled with intra-arterial infusion of an agonist is often useful for investigating the pharmacological effects of a compound or activation of a receptor subtype and the potential for interactions, physiological or pharmacological. However, studies with receptor antagonists often provide more valuable information concerning the physiology and pathophysiology of pathways, particularly when comparisons are made between patient and control groups. In this respect, the forearm vascular bed provides an ideal model with which to safely evaluate potential antagonists and investigate their effect on resistance vessels. Indeed, endothelin , angiotensin II , bradykinin  and substance P  antagonists, and both α- and β-adrenoceptor antagonists [67, 70] have been infused intra-arterially and provide valuable information about the contribution that such systems make to vascular tone at rest (Table 2), or during various manoeuvres such as activation of the sympathetic nervous system. Similarly, infusion of drugs that inhibit the in vivo formation of vasoactive mediators such as nitric oxide  or endothelin  can also provide useful information regarding the physiological importance of such systems. The forearm circulation can also be used to conduct agonist/antagonist interaction studies to assess the efficacy of a particular antagonist, which may be given either by local intra-arterial infusion [38, 69], or systemically . Once again, the local effects of antagonist in the forearm resistance bed correlate well with systemic effects [51, 69, 71–74].
Table 2. Systems affecting basal vascular tone in the human forearm.
Only when activated, does the renin-angiotensin system influence basal vascular tone, such as in sodium-depleted or volume-depleted individuals.
By inserting venous cannulae into both antecubital fossae, it is possible to obtain venous effluent from the infused and non-infused arm. This can provide useful information regarding local drug concentrations and the effect of the infused drug on other vasoactive mediators, the fibrinolytic system (see later), and platelet function. Data from the non-infused arm, serve as a contemporaneous control, and also provides information regarding systemic effects or drug ‘spill over’.
Considerations with intra-arterial infusions
The speed with which drugs and peptides act is an important consideration when designing experiments. Some drugs, such as 5-hydroxytryptamine, have a rapid onset of action with a short-lived maximal effect and, therefore, delaying assessment of FBF after starting an intra-arterial infusion may lead to an underestimation the maximum vascular response . In contrast, endothelin has a slow onset of action, maximum vasoconstriction being reached after ∼45 min . This effect was not investigated in the first intra-arterial studies , and a failure to consider the effects of prolonged infusion of endothelin appears to have led to excessive doses being given subsequently with serious adverse consequences . In addition, a slow onset of action makes it difficult to construct dose–response curves during a single study, which, as discussed, is the most powerful way of using the FBF technique. Similarly, tachyphylaxis may occur with repeated exposure to some agents, such as bradykinin  and substance P . In such circumstances it becomes necessary to compare responses on different occasions, which, because of the increased variability of measurement, may require a larger number of subjects to be studied, or for comparisons between active and placebo phases to be made. Therefore, a range of preliminary pilot studies may be required to assess the characteristics of the response to intra-arterial infusions before embarking on full dose ranging studies.
When making comparisons between groups, especially control and disease states, differences in basal blood flow or mean arterial pressure are potential sources of difficulty. A reduced response to an infused drug might be predicted when resting blood flow is higher, due to the effect of dilution and a reduced local drug concentration, although this should only become an issue at the extremes of flow due to the logarithmic nature of most dose–response curves. However, observational data indicate an increased absolute response to dilators and constrictors in subjects with higher basal flows [79, 80], but that the percentage change from baseline appears to be reasonably independent of baseline flow , which may, therefore, be a more suitable method for data analysis. However, which method of data analysis is more appropriate remains unclear. Indeed, different approaches are liable to produce conflicting results. Therefore, perhaps a more rigorous approach, as discussed below, is the use of appropriate ‘control’ agent(s).
Similarly, arterial pressure influences resting smooth muscle tone and thus the wall-to-lumen ratio, a higher mean arterial pressure, by virtue of an increased wall-to-lumen ratio, should theoretically enhance the response to vasodilators and vasoconstrictors . This may well explain the increased response to a wide-range of agents including noradrenaline and verapamil in subjects with hypertension [83–85]. Increasing resting tone reduces the in vitro response to vasodilators  but, clearly, low levels of resting tone may limit the degree to which vasodilatation can be detected and, therefore, explain why preconstriction can increase the response to some vasodilators [87, 88]. A similar problem exists during co-infusion of two vasoactive drugs, when changes in baseline flow produced by one may alter the response to the other. In such circumstances, use of FBF ‘clamp’ studies, as discussed in more detail below, may provide a better method with which to remove the potential influence of changes in baseline flow, but use of appropriate control drug(s) is an alternative approach.
We believe that the best way to account for differences in basal flow or arterial pressure between groups is the use of appropriate controls. For example, the inclusion of a ‘control’ vasoconstrictor, such as noradrenaline, in a protocol designed to investigate differences in the response to drugs, which are vasoconstrictors, between two groups (normotensive and hypertensive subjects, for example) may indicate whether any disparity is specific to that agent or is a general phenomenon, perhaps accounted for by baseline differences [52, 89]. The potential problem resulting from infusion of a drug which alters baseline FBF, such as NG-monomethyl-l-arginine (LNMMA), prior to administration of a second vasoactive agents, such as acetylcholine, can be tackled in a similar manner by using noradrenaline as a control vasoconstrictor . Indeed, occasionally it may be appropriate to use a control constrictor and a control vasodilator drug . However, now that ‘on-line’ measurement of FBF is possible, an alternative strategy is to adjust blood flow back to baseline after infusion of the first drug with an opposing agent (usually one that acts directly on smooth muscle), and then to administer the second drug. Such a system has already been use to investigate the effects of inhibition of endothelium derived nitric oxide production on the response to various drugs – the so-called ‘nitric oxide clamp’[91, 92].
Assessment of endothelial function
The vascular endothelium releases a number of biologically active mediators, including nitric oxide and endothelin-1, which regulate vessel tone and have been the focus of much research interest in recent years. It is now apparent that vasomotor regulation is only one aspect of endothelial function. Vasoactive mediators released by the vascular endothelium, including nitric oxide and endothelin-1 also have a number of other important actions including modulation of platelet aggregation and adhesion, smooth muscle cell growth, and leucocyte adhesion . In addition, the endothelium also synthesizes and releases a number of other substances including tissue plasminogen activator (tPA) [93, 94], a key enzyme in the fibrinolytic pathway, release of which can be potentiated by a variety of stimuli [95–97]. This suggests that the vascular endothelium not only plays a crucial role in regulating vessel tone and blood pressure, but also in preventing the development of atheroma, platelet activation and thrombus formation. Therefore, assessment of endothelial function has been the subject of much attention, most frequently as determined by vasomotor responses in the forearm, but techniques for measuring endothelial tPA release from the forearm vascular bed have been developed more recently .
Stimulated nitric oxide release
Endothelial function is commonly assessed by determining the vasodilator response, in the forearm, to brachial artery infusion of an agonist, which induces endothelial nitric oxide release via stimulation of nitric oxide synthase (endothelium-dependent agonist). It is usual also to assess the response to a direct nitric oxide donor drug, like sodium nitroprusside (endothelium-independent agonist), in order to exclude any concomitant alteration in vascular smooth muscle sensitivity to nitric oxide (Figure 7). A number of endothelium-dependent agonists have been employed in studies including: acetylcholine , bradykinin , 5-hydroxytryptamine  and substance P , although some, such as methacholine, appear to cause vasodilatation that is not nitric oxide mediated [101, 102]. Such agonists not only induce release of nitric oxide but also other vasodilator substances including endothelium derived hyperpolarizing factor (EDHF) and prostaglandins. Moreover, the nitric oxide-dependence varies between agonist [65, 88, 90], on the dose of agonist used , and between vascular beds for the same agonist . Of particular importance when acetylcholine is employed are differences in forearm length, since this is rapidly metabolized in vivo. Indeed, following brachial artery administration less than 1% of the infused acetylcholine reaches the hand , and the vasodilator response to acetylcholine is inversely related to forearm length . However, correction for forearm length is possible and, interestingly, reduces the apparent sex-related difference in the response to acetylcholine between men and women .
Using such methodology, endothelial dysfunction has been described in association with a number of cardiovascular risk factors (Table 3) including hypercholesterolaemia , diabetes mellitus , cigarette smoking  and ageing , although the results in hypertensive subjects are somewhat conflicting [47, 109–112]. Interestingly, endothelial dysfunction can be reversed by HMG CoA reductase inhibitors in hypercholesterolaemic individuals , and by angiotensin converting enzyme inhibitors in patients with hypertension , diabetes  and hypercholesterolaemia  (Table 4).
Table 3. Endothelial dysfunction in the human forearm vascular bed and cardiovascular risk factors.
Risk factor (number of subjects)
ACh – acetylcholine, MCh – methacholine, SNP – sodium nitroprusside, GTN – glyceryl trinitrate, NA – noradrenaline, LNMMA –NG-monomethyl-l-arginine.
Endothelial dysfunction appears to be a systemic phenomenon, affecting resistance and conduit vessels in the forearm as well as the coronary circulation. Recent data suggest that coronary endothelial dysfunction predicts long-term atherosclerotic disease progression and cardiovascular event rates [117, 118]. Interestingly, there is a correlation between forearm vasomotor responses and those in the coronary arteries [104, 119] and, therefore, the forearm vascular bed can probably be used as a surrogate for assessing endothelial function in the coronary arteries, minimizing the invasive nature of such investigations. However, forearm responses provide information concerning resistance vessels, whilst most coronary studies involve the epicardial arteries, which are conduit vessels. Moreover, in the presence of endothelial dysfunction acetylcholine may cause paradoxical coronary vasoconstriction, whereas reduced vasodilator effects typify the response in the forearm circulation. Nevertheless, endothelial dysfunction appears to occur early in the course of cardiovascular disease, and thus assessment of endothelial function may provide a mechanism for identifying those at increased risk, before they develop manifest vascular disease . Not only does this bring the promise of improved risk stratification, but also the possibility of targeting therapies designed to reduce risk factors or improve endothelial function, at those most in need.
Basal nitric oxide release
Basal release of nitric oxide in the forearm can be assessed by infusing the specific l-arginine analogue nitric oxide synthase inhibitor LNMMA  into the brachial artery. In healthy subjects, under resting conditions, LNMMA produces a dose-dependent reduction in FBF of up to 50% , which suggest that there is basal release of nitric oxide that opposes vasoconstrictor forces in the forearm. Interestingly, basal synthesis of nitric oxide may be higher in pre-menopausal women than men , and reduced in some conditions associated with endothelial dysfunction such as diabetes mellitus [122, 123], but not in hypercholesterolaemia [29, 124]. Again, in hypertension both reduced [89, 125] and normal  basal nitric oxide production have been described.
Co-infusion of LNMMA can also be used to assess the contribution of nitric oxide to any vasodilatation induced by a particular agonist, thus providing evidence of ‘endothelium-dependence’. However, co-infusion can also be used to address the important issue of up-regulation of alternative vasodilator pathways, such as endothelial-derived prostanoids, in conditions of endothelial dysfunction. For example reduced vasodilator response to carbachol (endothelium-dependent), and reduced vasoconstriction to LNMMA, has been described in microalbuminuric diabetics compared with controls . However, when carbachol was co-infused with LNMMA, dilator responses were reduced in the control group but not in the diabetics, suggesting that there was up-regulation of non-nitric oxide dilator pathways in the diabetic group. Similar results have been reported in individuals with hypertension  and hypercholesterolaemia .
Since LNMMA has a vasoconstrictor action, changes in baseline FBF may alter the response to subsequent agonists, in a non-specific manner. Likewise, a reduced response to infused LNMMA could also be a non-specific effect, and may also be seen for other ‘control’ vasoconstrictors . Therefore, the use of some form of control in studies using LNMMA is essential. This may involve using either a control dilator or vasoconstrictor,  or the ‘nitric oxide clamp’.
Other methods for assessing nitric oxide-dependent pathways
Endothelial function can also be assessed in the forearm by induction of ischaemia, which is the subject of a more detailed review in this series . This method was first suggested by Sinoway in 1989 , but introduced as the technique of flow-mediated dilatation by Celermajer et al. in 1992 . It employs forearm ischaemia as a stimulus to increase flow in the brachial artery, resulting in an increase in shear stress and, ultimately, nitric oxide-mediated dilatation of the brachial artery, which can be assessed using ultrasound. Sublingual GTN is frequently used as a control endothelium-independent dilator, although the precise dose seems to be a matter of some debate . While this technique has the advantage of being non-invasive, it requires considerable expertise and expensive equipment, and doubts still exist over its reproducibility  and validity. Moreover, studies demonstrating a close relationship between flow-mediated responses in the forearm and coronary endothelial function  are difficult to interpret since different stimuli (e.g. ischaemia and acetylcholine) are usually employed in the two beds. However, one may consider flow-mediated dilatation and intra-arterial studies as complementary, since they are assessing endothelial function in different parts of the arterial tree: flow-mediated dilatation provides information about arterial conduit vessels, and FBF about endothelial function in the resistance vessels.
Other aspects of vascular endothelial function
The vascular endothelium also plays an important role in the regulation of fibrinolysis. The endothelium synthesizes and releases tPA [93, 94], a key enzyme in the fibrinolytic pathway, which can be potentiated by a variety of stimuli [95, 96]. Using the basic technique of FBF coupled with intra-arterial drug infusion, Newby et al. have demonstrated that substance P increases plasma tPA in the infused arm compared with the non-infused arm (by assessing tPA activity in venous effluent) , and that this phenomenon is not due to increased blood flow per se. Moreover, the effect of substance P can be inhibited by co-administration of LNMMA . Interestingly, reduced stimulated tPA release has been described in cigarette smokers , which may provide a mechanism by whereby endothelial dysfunction may increase the risk of atherothrombosis due to reduced fibrinolytic capacity. This has also been confirmed in the coronary circulation recently .
There is increasing interest in the influence of genotype on the effect of drugs in vivo. However, drugs can also be used to characterize phenotypic differences between individuals with various genetic polymorphisms, which may or may not be apparent in the resting state. For example, the degree of vasodilatation in the forearm in response to an intra-arterial infusion of isoprenaline, a β-adrenoceptor agonist, is dependent on polymorphisms of the β2-adrenoceptor gene . Interestingly, the same polymorphisms also influence the response to inhaled β2-adrenoceptor agonists in patients with asthma . Clearly, therefore, assessing the FBF response to drugs given intra-arterially may provide a useful tool for improved phenotypic characterization of individuals and a better understanding of the potential clinical relevance of genetic polymorphisms.
Venous occlusion plethysmography provides a simple, and robust method for assessing blood flow in vivo. It is most frequently applied to the forearm and is a versatile technique that has proved extremely valuable in assessing human vascular physiology and pharmacology, especially when coupled with intra-arterial drug infusion. It has also been employed extensively to assess endothelial function in vivo, for which it remains the ‘gold-standard’. Unfortunately, the invasive nature of the technique has prevented its use in addressing the important question of whether changes in endothelial function, particularly following drug therapy, are a reliable surrogate of cardiovascular risk . Regrettably, flow-mediated dilatation is unlikely to be useful in this context either, and we must await a simple, reliable, non-invasive test of endothelial function to address this issue. However, plethysmographic assessment of FBF does provide an ideal method for assessing the effects of various drugs and endogenous peptides on the peripheral resistance vessels in vivo, without the need to conduct systemic studies. The technique is at its most powerful when used to compare responses within an individual during a single study, for example construction of dose–response curves before and after administration of an antagonist. Use of receptor antagonists has also extended our understanding of the physiological and pathological processes regulating blood flow in vivo. Clearly, almost 100 years after its first description, the place of venous occlusion plethysmography in physiology and pharmacology remains firmly established due to the robust validation of the technique and its great utility for addressing the important issues of the time.
We would like to thank Dr John Cockcroft for critical appraisal of the manuscript and helpful suggestions regarding key references, Professor Patrick Vallance for providing the information regarding experience of intra-arterial cannulation and Fiona Strachan for help with Figure 1.