Postgrad Med J: first published as 10.1136/pgmj.69.812.419 on 1 June 1993. Downloaded from Postgrad Med J (1993) 69, 419-428 © The Fellowship of Postgraduate Medicine, 1993 Review Article Microvascular investigations in mellitus Sandra J. Chittenden and Shukri K. Shami Department ofSurgery, University College and Middlesex School ofMedicine, London, UK

Summary: This paper reviews the current literature concerning the different investigative modalities available to assess the microcirculation in diabetic . The advantages and disadvantages of the different invasive and noninvasive methods available are presented objectively. We have concentrated on the tests that provide a quantitative assessment of the microcirculation, including laser Doppler fluxmetry, capillary microscopy, plethysmography, transcutaneous oximetry and radioactive isotope clearance. Some of the invasive methods described are now being replaced by noninvasive equivalents, providing similar information with less discomfort and risk to the patient.

Introduction The complications resulting from diabetic angio- information about the nature of the microcir- pathy can be severe and debilitating. An enormous culatory fault in diabetic may help in amount of research has been undertaken to to try understanding its aetiology and in developing new by copyright. clarify the mechanisms by which microangiopathy treatment strategies. Many of the invasive techni- occurs but has so far yielded no clear answer. The ques for microvascular measurement have been development of microangiopathy is probably superseded by noninvasive techniques, which is multi-factorial in origin and includes genetic sus- clearly advantageous for the investigation ofmicro- ceptibility. Factors that are thought to contribute angiopathy in the diabetic patient. This review towards the formation ofmicroangiopathy include places more emphasis on the noninvasive techni- those shown in Table I(a).' ques available. Unfortunately, by the time microvascular chan- become clinically evident, there is often little macrovascular assessment ges Importance of http://pmj.bmj.com/ that can be done in the way of treatment. It would be advantageous to identify those patients at risk of Care must be used in the interpretation of micro- developing severe retinal and glomerular impair- vascular assessments if macro- and microvascular ment and neuropathic or atherosclerotic lesions. disease coexist, and this is often the case, diabetic Functional microvascular changes may occur early patients being more prone to cardiovascular and on in the development of clinical angiopathy and peripheral arterial disease than non-diabetics.5'6 investigation of the microcirculation might there- The reasons for diabetics being more prone to fore provide a means of detecting angiopathy macroangiopathy are not certain but probably on September 28, 2021 by guest. Protected before it becomes clinically evident. Microvascular include the factors shown in Table I(b).5'7-'5 Oc- measurement techniques are also used clinically in clusive macrovascular disease may impede the the assessment of sympathetic neuropathy2 and ability ofthe microcirculation to respond to a given investigation of the peripheral microcirculation stimulus. It is therefore important to evaluate the may provide a way of detecting neuropathy in its extent of one source of circulatory dysfunction in earliest stages.3'4 order to study the other. In research, microcirculatory investigations might be used to provide an objective test of Microvascular investigations different modes of treatment. In addition, any An investigation of vessel morphology may be informative about angiopathy in its later stages and can help to assess its progression. However, in Correspondence: S.K. Shami, F.R.C.S., M.S., Surgical order to identify microcirculatory problems earlier Studies, Jules Thorn Building, Middlesex Hospital, on in their development, a functional assessment of Mortimer Street, London W1N 8AA, UK. the microcirculation may be more useful. This Accepted: 30 November 1992 usually consists of measuring blood flow, either in Postgrad Med J: first published as 10.1136/pgmj.69.812.419 on 1 June 1993. Downloaded from 420 S.J. CHITTENDEN & S.K. SHAMI

Table I Factors of importance in the pathogenesis of diabetic micro- and macroangiopathy (a) Microangiopathy Hyperglycaemia Results in nonenzymic glycosylation of proteins in the vessel walls -immune complexes Stimulate monocytes and macrophages to produce procoagulant activity' Platelet activation Initiates thrombosis, release of angiogenic factors such as PDGF and the transforming growth factors (TGF)-alpha and TGF-beta, and the release of vasoconstrictor substances such as thromboxane A2 Endothelial activation Results in increased permeability and release of vasoactive substances Macrophage and monocyte Produces free radicals (superoxides, hydroxyl radicals and peroxides) and activation cytokines and causes vascular damage (b) Macroangiopathy Hyperglycaemia A risk factor in atherogenesis5 which increases adherence of monocytes and macrophages to the vessel walls7'8 and may result in local accumulation of vasoactive substances, increased proliferation of endothelium due to the release of growth factors, and damage to the vessel wall by free radicals Free radicals Increased free radical activity is seen in diabetes9"0 and may be involved in microvascular damage"'12 Lipoproteins Concentrations of low density lipoproteins and very low density lipo- proteins are high in patients with poorly controlled diabetes1314 and this may be important in atherogenesis. High density lipoproteins are said to have a protective role in atherogenesis and are lower in some diabetics'5 Hypertension Not considered a 'very important factor in vascular disease in diabetes' by most authors Other factors Including: hyperinsulinaemia, rheological factors, growth factors, reduced physical activity, obesity and genetic factors PDGF = platelet-derived growth factor. by copyright. the steady state or in response to some provoca- entrainment studies suggest that the central and tion. The resting skin blood flow in diabetic local regulatory mechanisms provide coarse and patients is often shown to be normal and yet the fine microvascular control respectively.26 It is maximal flows are impaired.16"7 Since there has therefore important to investigate the different been some difficulty in finding abnormal data at systems separately, using both contralateral and rest (possibly because of some redundancy in the local stimuli. It is also desirable to study flow in control oflocal flow'8) many studies use some kind both the arterio-venous shunt vessels and the ofprovocation in order to expose a latent problem, nutritional capillaries in the skin. Unfortunately, http://pmj.bmj.com/ for example, looking at blood flow either during most studies to date are unable to distinguish exercise, while raising body temperature, after between flow in these different vessels. administering vasoactive substances'9 or following It may be meaningful to study the effects of tissue ischaemia. The actual pattern of the blood microcirculatory changes on the chemical milieu of flow response to a challenge such as ischaemia may the tissues, including oxygen tension, to determine be important.2 In diabetes mellitus, return of whether observed microvascular alterations have blood flow after ischaemia is impaired in the skin of any functional significance. Nuclear magnetic on September 28, 2021 by guest. Protected the toes21 and in the nailfold22 (although it occurs resonance (NMR) spectroscopy may be employed more rapidly in leg muscle'6). The cold recovery to study this by examining the redox status within time of skin blood flow is also longer.23 the tissues. One approach, which has not been Another way to study the function of the micro- exploited as yet, might be to study the fluorescence circulation is to evaluate reflexes such as the of the mitochondria as their metabolic state reduction of blood flow normally seen on standing changes. (the veno-arteriolar reflex). This is impaired in patients with .24 Similarly, in the retina, the vasoconstriction normally caused by Investigative techniques high oxygen tension is blocked by hyperglycae- mia.'6'25 Care must be taken in interpreting the Blood tests and other nonspecific tests changes in blood flow induced by such stimuli since there are various controlling influences that may be It would be useful if a simple clinical examination at work. These include sympathetic nerve activity, or a laboratory test carried out on the blood or local autoregulatory mechanisms, venous reflexes urine samples routinely taken from diabetic pa- and the status of arterio-venous shunts. Thermal tients could show or predict the development of MICROVASCULAR INVESTIGATIONS IN DIABETES 421 Postgrad Med J: first published as 10.1136/pgmj.69.812.419 on 1 June 1993. Downloaded from angiopathy. It is known that hyperglycaemia has at order to study capillary permeability and blood- least a permissive role in the aetiology of diabetic tissue exchange, it is usually necessary to introduce lesions, such as retinopathy and nephropathy.27 labelled tracer substances invasively. However, Monitoring the levels of glucose or glycosylated their movement can now be monitored non- haemoglobin in the blood and attempting optimal invasively, so reducing the trauma presented by the glycaemic control may therefore help in preventing investigation. the development of microvascular complications. However, the relationship between these factors is Histology Histological tissue measurement tech- not strict and metabolic control is oflittle use once niques show that blood vessel walls are thicker in microangiopathies have developed. One of the diabeties mellitus and vascular lumina are nar- standard tests for diabetic microangiopathy is rower, especially in those patients with vascular microalbuminuria, resulting from glomerular dys- complications.33 Capillary basement membrane function. Early elevation of transcapillary leakage thickening has been said to be the hallmark of ofalbumin in the kidney (incipient nephropathy) is microvascular disease.34 However, it occurs in said to be the first sign ofmicrovascular disease.28 If parallel with microangiopathy, rather than being a performed regularly in all diabetic patients, this test cause, and tissue biopsy for investigation of base- might facilitate the detection of angiopathy in its ment membrane thickness is a poor substitute for a earliest clinical stages, but there are obvious finan- functional assessment of the microcirculation.35 cial and practical limitations. Another marker for angiopathy is a slight increase in blood pressure, although this is not very specific. An increased level Local clearance (bloodflow) Clearance of radio- of plasma von Willebrand factor has been pro- active isotopes such as '33Xe and Nal31I can be used posed as an indicator of microangiopathy since its to measure perfusion. By injecting '33Xe into mus- elevation is associated with retinal hyperperme- cle36'37 or skin38 its rate of disappearance can be ability.29 Angiotensin converting enzyme (which used to measure local blood flow. 133Xe clearance is cleaves I to II and inac- thought to reflect blood flow since it is freely

angiotensin angiotensin by copyright. tivates bradykinin) is also high in some diabetic diffusible and fat soluble and crosses all vascular patients and these have a higher incidence of barriers. In contrast, the clearance ofNa'31I may be retinopathy. Angiotensin converting enzyme was limited by capillary permeability rather than blood therefore thought to indicate widespread endo- flow. This means that in diabetes, where capillary thelial damage in microvascular disease.30 How- thickening is often seen, rates ofblood flow may be ever, it is increased in only a small percentage of underestimated by Nal31I clearance rates alone.39 patients with microangiopathy and is a less reliable Using combined xenon and sodium clearance indicator than von Willebrand factor. Increased measurements it is possible to study vessel-tissue plasma levels ofinactive renin may also be a marker exchange as well as blood flow.37 for microvascular complications. Although in http://pmj.bmj.com/ diabetic patients without microvascular complica- Skin perfusion pressures In this technique, an tions, plasma inactive renin concentrations lie intradermal injection of a radioactive tracer mixed within the normal age-adjusted range, in those with histamine (to promote vasodilatation) is given patients with retinopathy or albuminuria it is and the washout measured with a scintillation nearly always above this range.31 Similarly, limited counter." A blood pressure cuffis applied over the joint motility in childhood diabetes mellitus indi- injection site and the pressure in it increased until cates an washout This cuff is taken as increased risk for microvascular disease.32 stops. pressure the on September 28, 2021 by guest. Protected The problem with many of these tests is that they skin perfusion pressure (SPP). A number of differ- indicate the presence of some process which ent radioactive tracers, including '31I-antipyrine and accompanies angiopathy, rather than measuring it 99mpertechnate, have been used. This technique has directly. They may therefore be useful as epidemio- its disadvantages, as it is time consuming, espe- logical pointers but are of little help in predicting cially if the SPP is to be determined at more than and following angiopathy in an individual patient. one site, and painful enough to require analgesia. An alternative new approach using either a Invasive techniques photoplethysmography transducer41 or a laser Doppler probe (see laser Dopplerfluxmetry below) Many of the microcirculatory changes in diabetic underneath an inflatable cuff may be more useful. microangiopathy were originally discovered using invasive techniques. Although in some cases these Capillary filtration and permeability Increased techniques have since been replaced by noninvasive capillary diffusion capacity (CDC) has been dem- methods, in others there is no substitute for the onstrated by various investigators in different invasive approach. Instead, the methods have been tissues (such as skeletal muscle,37'42'43 retina44 and made as atraumatic as possible. For example, in nerves45) for small ions,37'42 dopamine,46 fluor- 422 S.J. CHITTENDEN & S.K. SHAMI Postgrad Med J: first published as 10.1136/pgmj.69.812.419 on 1 June 1993. Downloaded from escein,29'47 albumin,48 IgG49 and 131I.50 Microvas- offlow. However, skin temperature has a nonlinear cular permeability is related to the metabolic state relationship with blood flow, especially at higher and may be reduced by metabolic regulation.5 The levels (above 28°C) and responds slowly to flow permeability of the vascular diffusion barriers, i.e. changes, lagging behind them.62 Skin temperature the capillary wall and pericapillary collagen sheath, may be useful for demonstrating large changes in can be measured using the 'relatively atraumatic the perfusion of skin after drug therapy63 or technique' of intravital fluorescence videomicro- surgery. The maximum increase in finger skin scopy,47 especially in the eye44'52 and nailfold.22 temperature following ischaemia is decreased in Sodium fluorescein passes the capillary wall and diabetes mellitus33 and is still less in those patients pericapillary space faster in diabetic patients. How- with vascular complications. ever, the results of such studies should be inter- Thermal clearance probes have been used to preted with caution, since there may be effects due measure blood flow noninvasively in the skin.64 The to, for example, the lymphatic return of the tracers probes consist typically of a heated central copper or to the labelled substances being handled disc with an unheated copper annulus. The tem- differently from unlabelled ones.53 perature difference between these is measured using thermocouples and decreases as the blood flow Measurement of capillary pressure using micro- beneath the probe increases.65 Although, in muscle, pipettes Capillary pressure can be measured the measurement depends greatly on the proximity directly using a microinjection technique.54'55 Using ofthe probe to a blood vessel,66 in the skin, thermal this method the pressure in the arteriolar end, the clearance probes measure total skin blood flow loop and the venular end of the capillary can be (including flow through arterio-venous anastomoses measured, and an assessment of the pressures in the microcirculation) and not nutritive flow responsible for fluid filtration and reabsorption only.67 The technique can give important inform- evaluated. Nailfold capillary pressure have been ation about flow distribution, especially if used in found to be elevated in young insulin-dependent conjunction with a measure of total flow, such as diabetics in the early stages of disease.5 The laser Doppler flowmetry.66 An alternative method regulation of pressure in relation to skin tempera- is to measure the power consumption of a thermo- by copyright. ture is also disturbed and the pressure during statically controlled heating element.68 The higher postocclusive reactive hyperaemia is less than in the blood flow, the more rapidly heat will be normal subjects.57 removed from the skin under the heating element and the more power will have to be supplied to Noninvasive techniques maintain a fixed temperature. This approach is particularly attractive when the heating element It is preferable that any routine investigations be used is that ofa tcPo2 electrode or thermostatically noninvasive. Some of the tests described above controlled laser Doppler probe holder. Oxygen have been abandoned since the information needed tensions and blood flow measurements have been http://pmj.bmj.com/ can be obtained using the noninvasive tests des- made simultaneously in this way.68 cribed below. Such tests are less traumatic to the patient and can be repeated as necessary to assess Venous occlusion plethysmography Various tech- the development and progression of angiopathy. niques of plethysmography have been used for the measurement of limb blood flow. The method is Vessel morphology Abnormalities in vessel mor- able to give quantitative values after calibration phology can be investigated using in vivo micro- and is said to be accurate and It reproducible.69 on September 28, 2021 by guest. Protected scopy as well as by histological methods. Dilated relies on measuring volume changes, the rate of nailfold capillaries have been found, and well increase in limb volume following release of occ- defined in diabetes mellitus.22'58 59 However, there is lusion being used to measure blood flow rates. The a difference in capillary morphology even between usual volumetric techniques are water displace- toes on the same foot60 and so care must be taken in ment70 or a mercury-in-Silastic strain-gauge.71 In making measurements. The first retinal lesions in the hand and foot, values from venous occlusion can be seen noninvasively by plethysmography are assumed to reflect skin blood fundoscopy or even earlier using fluorescein flow since there is little muscle present. However, in angiofluorography.6' the arm and leg, the technique cannot distinguish between skin and muscle flow (unless used in Skin temperature and thermal clearance The tem- conjunction with adrenaline ionophoresis66). In perature of the skin tends to increase as the blood normal subjects, under thermoneutral conditions, flow increases. It can be measured most simply the muscle and skin contribute similarly to total using thermocouples, although many studies use limb blood flow so that changes in total flow are thermography which is useful for showing gross assumed to reflect changes in skin and muscle flow changes, such as the cessation and recommencing equally, but this may not hold true for diabetic MICROVASCULAR INVESTIGATIONS IN DIABETES 423 Postgrad Med J: first published as 10.1136/pgmj.69.812.419 on 1 June 1993. Downloaded from patients.66 In addition, whilst a thermal stress in a nutritional capillaries alone. The dependence of non-diabetic patient results in changes in skin LDF readings on depth of penetration means that blood flow only, this may not be the case in care must be taken when comparing readings from diabetes. Additional problems with the technique diabetic patients in whom there may be epidermal are that alterations in tissue compliance occur with thickening, with normal subjects. Synchronous each pulse and that a veno-arteriolar reflex may be measurement using LDF and microscopy in mus- induced by cuff compression if it is high or cle77 and skin78 gives broadly comparable results prolonged.2 but not under all conditions.79 In the skin there are Strain gauge plethysmography has been used to differences in the timing of spontaneous oscilla- study blood flow and capillary diffusion coefficients tions, the time course of reactive hyperaemia and in the forearm. Following release of occlusion, the the response to venous hypertension between the initial slope of the volume/time curve represents two measurement techniques. In addition, LDF venous filling, and is used to calculate blood flow. cannot distinguish between capillary flow and Vascular resistance and venous capacity can also shunt flow and it is therefore now regarded as a be calculated. The shallower slope that follows measure oftotal blood flow in the skin.78 LDF may after a few minutes represents capillary filtration. be a particularly useful technique in diabetes where This method is simple, quick and noninvasive. there is thought to be increased non-nutritive flow in the skin due to sympathetic denervation.66 Photoelectric plethysmography Using this techni- The laser Doppler flowmeter does not provide que, light is shone onto the skin where it is absorbed absolute quantitative values relating to blood flow by both skin and blood.73 As the volume ofblood in and calibration is difficult. The measurements show the skin increases, it absorbs more light and less is marked variations over distances of less than received by the detector. The method cannot 1-2 cm66 and readings are generally regarded as distinguish between nutritive and shunt flow and is not repeatable, although one series of measure- sensitive to movement and to the orientation and ments demonstrated a coefficient of variation of ofthe red cells. Ifthe source used also less than 12%.80 it be valid to packing light Although may by copyright. produces heat, it may cause vasodilatation which examine the relative changes in flow within an will interfere with the measurements. It can individual subject, comparisons of absolute blood therefore only be used for qualitative comparisons. flow rates between groups are of doubtful validity, The technique has now been largely superseded by especially when there are changes which may affect other methods. the laser Doppler signal (such as the glycosylation of connective tissues or thickening of capillary Laser Doppler fluxmetry basement membranes seen in diabetic angiopathy). In spite ofits limitations, LDF is better validated Many observations of blood flow, in the as a measure of superficial microvascular volume especially http://pmj.bmj.com/ skin, have been made using laser Doppler flow- flow rate than its predecessor photoelectric plethys- metry (LDF). This involves shining a laser light mography.81 Using LDF it has been demonstrated into the skin and measuring the back scattered that basal blood flow in the skin is not altered in light.74 Some ofthe incident light will strike moving diabetes mellitus but that maximal flows are red blood cells and be reflected with a shift in its impaired.8 Blood flow has also been measured in frequency, caused by Doppler broadening, which is the retinal arterioles in diabetes mellitus using related to the speed of the cells. Studies in vitro LDF.83 show that the mean Doppler frequency is propor- on September 28, 2021 by guest. Protected tional to the rate ofblood flow.74,75 In vivo studies in Capillary microscopy Microscopy facilitates the skin using '3Xe washout techniques confirm this direct and noninvasive study of the capillaries in linearity75'76 and show that stable recordings can be the upper layers of the skin84 and in this way achieved.76 A laser Doppler flowmeter provides a morphological abnormalities can be observed. In signal, used as a measure of blood flow, which is addition to such static investigations, blood flow in proportional to the number of red cells in the the capillaries can be measured. Some work has volume penetrated by the laser beam and to the involved tracers such as fluorescein to measure integrated red cell velocity. The laser beam pene- retinal blood flow.61 In other studies, the capillary trates through the skin and measures flow in a blood velocity itself is measured84'85 and along with volume of several cubic millimeters. Although the capillary morphometry and estimation of the beam is supposed to reach a depth of 0.6 mm,74 it relative haematocrit, this allows volume flow pat- maintains halfofits sensitivity at a depth of 1.2 mm terns to be studied.86'87 Because the capillaries are and can still pick up some cell movement deeper under direct observation, microscopy is the only than this.66 This means that flow in small arterioles technique in which it is certain that nutritive blood and arterio-venous anastomoses may be included flow alone is being measured. Simultaneous LDF so that LDF does not give a measure offlow in the and capillary blood velocity measurements have 424 S.J. CHITTENDEN & S.K. SHAMI Postgrad Med J: first published as 10.1136/pgmj.69.812.419 on 1 June 1993. Downloaded from been used to demonstrate discrepancies between capillary blood flow against which other techniques total and nutritional flow in skin.77'78 must be judged.78 Capillary microscopy has been Early studies used time-consuming frame-by- used to demonstrate dilated capillaries and delayed frame analysis of capillary images recorded on return of local blood flow after ischaemia in video tape.88 This involves locating the position ofa diabetes mellitus.22 cell group or plasma gap, advancing the tape by several frames and measuring the new position of Noninvasive measurement of capillary filtration the cells and gaps. The velocity can be calculated The invasive measurement of capillary filtration from the distance moved in a set time (which is the coefficients (CFCs) has been described above. number of frames multiplied by 1/25 or 1/30 Capillary filtration can also be measured non- seconds). In the simpler 'flying spot' technique,89 invasively by venous occlusion plethysmography the speed of a moving spot superimposed on the (described above). The slower part of the volume/ video image is adjusted until it moves at the same time curve represents extravasation of plasma into rate as the red cells. The analysis can be carried out the tissues as a result of elevated venous pressure. in real time but is subject to user-error and cannot The CFC is normal or depressed early on in detect rapid changes in velocity. diabetes92 but increases with duration and becomes More sophisticated techniques have also been elevated in long-standing diabetes.93 Although part developed, based on early photodetector meth- of this increase may be an artefact,53 due to ods.90 They rely on the principles of cross-cor- impairment ofthe veno-arteriolar reflex in diabetic relation, in either the space84 or the time85 domains, patients,94 such effects are not entirely responsible to match up the patterns created as dark red cells for increased filtration94 and so there is a real and light plasma gaps pass the photodetectors (or a increase in the CFC in long-duration diabetes. window in the video image). Subject movement presents a significant problem Noninvasive measurement of xenon clearance As when these techniques are applied in humans. Ifthe has been discussed, clearance of the radioactive area being studied moves even slightly underneath isotope '33Xe has been used to measure blood flow the microscope objective, then the capillaries will in the skin.38 In order to avoid the initial hyper- by copyright. move within the video image. Attempts have been aemia due to injection trauma, 33Xe can also be made to immobilize the area under study using introduced epicutaneously via a small chamber special casts or brackets9l but these do not always attached to the skin.95 Since its molecular weight completely eliminate movement and may interfere and solubility are similar to those of oxygen, 133Xe with blood flow. Newer systems compensate for the uptake has been used as an indicator of vascular movement by either tracking a capillary as it moves permeability in assessing the oxygen diffusion around in the image,86'87 or by performing a barrier in chronic venous disease.95 two-dimensional cross-correlation on a small area of the image ('CapiFlow system for the evaluation Transcutaneous oxygen tension Transcutaneous http://pmj.bmj.com/ of video recorded dynamic capillary blood flow oxygen pressures have been used to study the end parameters', CapiFlow AB, Sweden). Using such result of perfusion and diffusion effects in diabetes tracking methods, velocimetry can be carried out in mellitus by monitoring tissue oxygenation. Some real time on capillary microscopy images in which studies show that the transcutaneous pressure of there is subject movement. The technique is still oxygen (tcPo2) is slightly lower in patients with time consuming, especially when several capillaries diabetes mellitus than in normal subjects.96 How- are but is much less so than earlier it is not clear how well skin studied, ever, oxygen pressures on September 28, 2021 by guest. Protected methods. reflect the normal levels of oxygen in the tissues. Unfortunately the technique of velocimetry The technique ofmeasuring tcPo2 was originally using noninvasive capillary microscopy is only developed for monitoring arterial oxygenation in applicable to certain sites such as the nailfold and neonates. In order to make these measurements, it only a small sample ofcapillaries can be studied at a is necessary to 'arterialize' the skin by heating it to time. The velocity of blood flow in the different 43 or 44°C.97 This produces a local hyperaemia, so capillaries may differ substantially. Since there are that excess oxygen diffuses across the skin to a differences in capillary morphology even between modified Clarke electrode in the tcPo2 probe where the toes of the same foot,60 it is important to it is chemically reduced and measured. TcPo2 standardize the site of measurement. Capillary measured using a heated probe thereafter does not microscopy does, however, permit characterization show actual tissue oxygenation under normal phys- of the size, shape and number of vessels present so iological conditions. When arterial Po2 is adequate, that the influence of these parameters on blood tcPo2 changes with blood flow and when blood flow flow can be studied. is adequate, tcPo2 reflects the arterial oxygen In spite of its limitations, microscopy is still saturation. regarded as the 'gold standard' for measuring Heating the skin to 44°C produces a 20-fold Postgrad Med J: first published as 10.1136/pgmj.69.812.419 on 1 June 1993. Downloaded from MICROVASCULAR INVESTIGATIONS IN DIABETES 425 increase in blood flow in diabetic patients com- the same9 and such functional studies are able to pared with a 40-fold increase in normal subjects as show differences between diabetic patients and measured by LDF.24 Care must therefore be taken controls before there are any clinical or mor- in choosing the electrode temperatures and in phological signs of microangiopathy. interpreting the results since any differences seen In measuring tcPo2, a control electrode is often may reflect an inability to increase the nutritional used, placed in the subclavicular area. The tcPo2 in blood flow at 44'C. When skin oxygen tension in the area of interest is then related to this, the ratio venous disease is measured at 37°C, there is no being used to give a regional perfusion index."° difference between patients and controls at rest,98 Such indices may provide better discrimination whereas there is at 44°C. The poor oxygenation of than single tcPo2 values. the tissues sometimes found using tcPo2 measure- tcPo2 readings are dependent on arterial and ments in diabetes mellitus63'96 may therefore be a venous blood pressures,'10 arterio-venous pressure result of the method of measurement rather than a differences'02 and changes in venous P02, and are true pathophysiological feature. Furthermore, al- linearly related to the blood flow under the elec- though the capacity for the microvessels to vaso- trode.68'103 Epidermal thickness also has an dilate and increase their flow to supply oxygen to influence, tcPo2 falling as the thickness increases.04 the tcPo2 probe may be reduced in diabetes, there is In addition, tcPo2 measurements depend on capil- no evidence that this has any functional lary density'05 and the number of perfused significance and patients with diabetes mellitus are superficial capillaries06 and may also be affected by probably still able to supply their tissues with angiopathy in these capillaries. Other factors have sufficient oxygen. also been shown to affect tcP02, including Although few studies actually show a inflammation, oedema and the skin's oxygen con- significantly lower tcPo2 in diabetics at rest, pos- sumption. Care must be therefore used when sibly because ofredundancy in control oflocal flow making comparisons of absolute tcPo2 levels, and metabolic accommodation to modest especialy when studying diabetic subjects, who are more hypoxia,18 pronounced differences can be known to differ in some of these variables. by copyright. evinced using an oxygen inhalation test.99 In Breuer tcPo2 may be better for measuring induced et al.'s study,96 although the basal tcPo2 is only responses such as postocclusive hyperaemia'07 than slightly lower in diabetics, there are more basal blood flow. However, great care must be significant differences when breathing 5 and 10 taken when using the heated tcPo2 to monitor litres of oxygen per minute. However, what is microvascular responses in this way, since it has actually being measured is the system's maximal been shown that heating the skin to 43 or 44°C capacity to supply oxygen under conditions of abolishes both periodic changes in blood flow and hyperaemia and hyperoxia. Nevertheless, the rate the normal blood flow regulation by local reflexes of rise of oxygen concentration is slower in or vasodilatory substances.68,'08 http://pmj.bmj.com/ diabetics, whilst the time to the maximum level is

References 1. Uchman, B., Bang, N.U. & Rathbun, M.J. Effect of insulin 7. Setiadi, H., Wautier, J., Courillon-Mallet, A. et al. Increased immune complexes on human blood monocyte and endo- adhesion to fibronectin and MO-1 expression by diabetic thelial cell procoagulant J Lab Clin Med 112: J Immunol 138: activity. 1988, monocytes. 1987, 3230-3234. on September 28, 2021 by guest. Protected 652-659. 8. Brownlee, M., Vlassara, H. & Cerami, A. The pathogenetic 2. Low, P.A., Neumann, C., Dyck, P.J. et al. Evaluation ofskin role of non-enzymatic glycosylation in diabetic complica- vasometer reflexes by using laser Doppler velocimetry. tions. In Crabbe, M.J.C. (ed.) Diabetic Complications: Mayo Clin Proc 1983, 58: 583-592. Scientific and Clinical Aspects. Churchill Livingstone, Edin- 3. Archer, A.G., Roberts, V.C. & Watkins, P.J. Blood flow burgh, 1987, pp. 94-139. patterns in painful diabetic neuropathy. Diabetologia 1984, 9. Jennings, P.E., Jones, A.F., Florkowski, C.M. et al. In- 27: 563-567. creased diene conjugates in diabetic subjects with microan- 4. Shore, A.C., Price, K.J., Tripp, J.H. et al. Functional giopathy. Diabetic Med 1987, 4: 452-456. assessment of the microcirculation in children with diabetes 10. Collier, A., Wilson, R., Bradley, H. et al. Free radical mellitus. Conference Proceedings. Joint Meeting of the activity in . Diabetic Med 1990, 7: 27-30. Italian andBritish Microcirculation Societies, Exeter Univer- 11. McCord, J.M. Oxygen-derived free radicals in post- sity, 1989. ischaemic tissue injury. N Engl J Med 1985, 312: 159-163. 5. Brand, F.N., Abbott, R.D. & Kannel, W.D. Diabetes, 12. Wolff, S.P. The potential role of oxidative stress in diabetes intermittent claudication, and risk ofcardiovascular events. and its complications: novel implications for theory and The Framingham study. Diabetes 1989, 38: 504-509. therapy. In: Crabbe, M.J.C. (ed.) Diabetic Complications: 6. Jonason, T. & Ringqvist, I. Diabetes mellitus and intermit- Scientific and Clinical Aspects. Churchill Livingstone, Edin- tent claudication. Relation between peripheral vascular brugh, 1987, pp. 167-220. complications and location of the occlusive atherosclerosis in the legs. Acta Med Scand 1985, 218: 217-221. 426 S.J. CHITTENDEN & S.K. SHAMI Postgrad Med J: first published as 10.1136/pgmj.69.812.419 on 1 June 1993. Downloaded from

13. Lopes-Virella, M.F., Wohltmann, H.J., Mayfield, R.K. et 35. Doran, T.L. & Tattersall, R.B. Blind alleys in diabetes al. Effect of metabolic control on lipid, lipoprotein and research: muscle capillary basement thickening: marker of apolipoprotein levels in 55 insulin-dependent diabetic microvascular complications or false prophet? Diabetic Med patients - a longitudinal study. Diabetes 1983, 32: 20-25. 1986, 3: 413-418. 14. Sosenko, J.M., Beshow, J.L., Meittinen, O.S. et al. Hyper- 36. Lassen, N.A., Lindbjerg, J. & Munck, O. Measurement of glycaemia and plasma lipid levels: a prospective study of blood-flow through skeletal muscle by intramuscular injec- young insulin-dependent diabetic patients. N Engl J Med tion of xenon-133. Lancet 1964, 1: 686-689. 1980, 302: 650-654. 37. Trap-Jensen, J., Alpert, J.S., del Rio, G. et al. Capillary 15. Feher, M.D., Stevens, J., Lant, A.F. et al. Importance of diffusion capacity for sodium in skeletal muscle in long-term routine measurement of HDL with total cholesterol in juvenile diabetes mellitus. Acta Med Scand (Suppl) 1967, diabetic patients. J R Soc Med 1991, 85: 8-11. 476: 135-146. 16. Christensen, N.J. Muscle blood flow, measured by xenon133 38. Moore, W.S. Determination of level: measure- and vascular calciferations in diabetics. Acta Med Scand ment ofskin blood flow with Xenon Xe 133. Arch Surg 1973, 1968, 183: 449-454. 107: 798-802. 17. Sigroth, K. Reflex vasodilatation of the fingers in the study 39. Roddie, I.C. Circulation to skin and adipose tissue. In: of peripheral vascular disorders: with special reference to Shepherd, J.T. & Abboud, F.M. (eds) Handbook ofPhysi- diabetes mellitus. Acta Med Scand Suppl. 1957, 325:1-116. ology. The cardiovascular system: peripheral circulation and 18. McMillan, D.E. The microcirculation: changes in diabetes organ blood flow. Section 2, Vol III, Part 1, Chapter 10. mellitus. Mayo Clin Proc 1988, 63: 517- 520. American Physiological Society, Bethesda, MD, 1983, 19. Weindorf, N,. Shultz-Ehrenburg, U. & Altmeyer, P. Diag- pp. 285-317. nostic assessment of diabetic microangiopathy by TcPO2 40. Holstein, P., Lund, P., Larsen, B. et al. Skin perfusion stimulation tests. Adv Exp Med Biol 1987, 220: 83-86. pressure measured as the external pressure required to stop 20. Jorgensen, R.G., Russo, L., Mattiol, L. et al. Early detection isotope washout: methodological considerations and the ofvascular dysfunction in . Diabetes 1988, 37: normal values on the legs. Scand J Clin Lab Invest 1977, 37: 292-296. 649-659. 21. Christensen, N.J. Spontaneous variations in resting blood 41. van den Broek, T.A.A., Dwars, B.J., Rauwerda, J.A. et al. flow, post ischaemic peak flow and vibratory perception in Photoplethysmographic selection of amputation level in the feet of diabetics. Diabetologia 1969, 5: 171-178. peripheral vascular disease. J Vasc Surg 1982, 8: 10-13. 22. Fagrell, B., Hermansson, I.-L., Karlander, S.-G. et al. Vital 42. Leinonen, H., Maitkainen, E. & Juntunen, J. Permeability capillary microscopy for assessment of skin viability and and morphology of skeletal muscle capillaries in type 1 microangiopathy in patients wtih diabetes mellitus. Acta (insulin-dependent) diabetes mellitus. Diabetologia 1982,22: Med Scand Suppl 1984, 687: 25-28. 158-162.

23. Hatanaka, H., Matsumoto, S., Kitamura, Y. et al. Skin 43. Alpert, J.S., Coffman, J.D., Balodimos, M.C. et al. Capillary by copyright. blood flow in diabetic patients during cold loading. Kobe J permeability and blood flow in skeletal muscle of patients Med Sci 1989, 35: 131-136. with diabetes mellitus and geneticprediabetes. NEnglJ Med 24. Rayman, G., Williams, S.A., Spencer, P.D. et al. Impaired 1972, 286: 454-460. microvascular hyperaemic response to minor skin trauma in 44. Krogsaa, B., Lund-Andersen, H., Mehlsen, J. et al. Blood- type I diabetes. Br Med J 1986, 292: 1295-1298. retinal barrier permeability versus diabetes duration and 25. Atherton, A., Hill, D.W., Keen, H. et al. The effect of acute retinal morphology in insulin dependent diabetic patients. hyperglycaemia on the retinal circulation of the normal cat. Acta Opthalmol Copenh 1987, 65: 686-692. Diabetologia 1980, 18: 233-237. 45. Rechthand, E., Smith, Q.R., Latker, C.H. et al. Altered 26. Lafferty, K., De Trafford, J.C., Roberts, V.C. et al. On the blood-nerve barrier permeability to small molecules in nature of Raynaud's phenomenon: the role of histamine. experimental diabetes mellitus. J Neuropath Exp Neurol Lancet 1983, 2: 313-315. 1987, 46: 302-314. 27. Tchobroutsky, G. Why do some diabetics develop severe 46. Lorenzi, M., Karam, J.H., Mcllroy, M.B. et al. Increased http://pmj.bmj.com/ microvascular complications? J Diabetic Complications growth hormone response to dopamine infusion in insulin 1989, 3: 1-5. dependent diabetic subjects. J Clin Invest 1980, 65: 146-153. 28. Viberti, G.C., Hill, R.D., Jarrett, R.J. et al. Micro- 47. Bollinger, A., Frey, J., Jager, K. et al. Patterns of diffusion albuminuria as a predictor of clinical nephropathy in through skin capillaries in patients with long-term diabetes. insulin-dependent diabetes mellitus. Lancet 1982, i: N Engl J Med 1982, 307: 1305-1310. 1430-1432. 48. Feldt-Rasmussen, B. Increased transcapillary escape rate of 29. Porta, M., Townsend, C., Clover, G.M. et al. Evidence for albumin in Type I (insulin-dependent) diabetic patients with functional endothelial cell damage in early diabetic microalbuminuria. Diabetalogia 1986, 29: 282-286.

retinopathy. Diabetologia 1981, 20: 597-601. 49. Parving, H.-H. & Rossing, N. Simultaneous determination on September 28, 2021 by guest. Protected 30. Lieberman, J. & Sastre, A. Serum angiotensin-converting of the transcapillary escape rate of albumin and IgG in enzyme: elevations in diabetes mellitus. Ann Intern Med normal and long-term juvenile diabetic subjects. Scand J 1980, 93: 825-826. Clin Lab 1973, 32: 239-244. 31. Luetscher, J.A., Kramer, F.B., Wilson, D.M. et al. Increased 50. Roztocil, K., Prerovsky, I. & Oliva, I. Capillary diffusion plasma inactive renin in diabetes mellitus indicates increased capacity for 1-131 in the subcutaneous tissue of the lower risk for microvascular disease. N Engl J Med 1981, 305: extremeties in patients with diabetes mellitus. Vasa 1976, 5: 191-194. 338-341. 32. Rosenbloom, A.L., Silverstein, H., Lezotte, D.C. et al. 51. Parving, H.-H., Noer, I., Deckert, T. et al. The effect of Limited joint motility in childhood diabetes mellitus metabolic regulation on microvascular permeability to small indicates increased risk for microvascular disease. N Engl J and large molecules in short-term diabetes. Diabetologia Med 1981, 305: 191-194. 1976, 12: 161-166. 33. Ajjam, Z.S., Barton, S., Corbett, M. et al. Quantitative 52. Waltman, S. Sequential vitreous fluorometry in diabetes evaluation of the dermal vasculature of diabetics. Q J Med mellitus: a marker of microvascular complications. Trans 1985, 54: 229-239. Am Ophthalmol Soc 1984, 82: 827-849. 34. Siperstein, M.D. Diabetic microangiopathy. Genetics, 53. Tooke, J.E. The microcirculation in diabetes. Diabetic Med environment, and treatment. Am J Med 1988, 85 1987, 4: 189-196. (Suppl 5A): 115-130. MICROVASCULAR INVESTIGATIONS IN DIABETES 427 Postgrad Med J: first published as 10.1136/pgmj.69.812.419 on 1 June 1993. Downloaded from

54. Landis, E.M. Microinjection studies of capillary blood 76. Stem, M.D., Lappe, D.L., Bowen, P.D. et al. Continuous pressure in human skin. Heart 1930, 15: 209-228. measurements oftissue blood flow bylaser-Doppler spectro- 55. Williams, S.A., Wasserman, S., Rawlinson, D.W. et al. scopy. Am J Physiol 1977, 232: H441-448. Dynamic measurement of human capillary blood pressure. 77. Tyml, K. & Ellis, C.G. Simultaneous assessment of red cell Clin Sci 1988, 74: 507-512. perfusion in skeletal muscle by laser Doppler flowmetry and 56. Shore, A.C., Sandeman, D.D. & Tooke, J.E. Nailfold videomicroscopy. Int J Microcirc Clin Exp 1985, 4: capillary pressure in non-nephropathic diabetic patients of 397-406. moderate disease duration: The effect ofangiotensin conver- 78. Tooke, J.E., Ostergren, J. & Fagrell, B. Synchronous ting enzyme inhibition by enalapril. British Microcirculation assessment ofhuman skin microcirculation by laser Doppler Society Meeting, April 1992, Birmingham University. flowmetry and dynamic capillaroscopy. Int J Microcirc Clin 57. Tooke, J.E. A capillary pressure disturbance in young Exp 1983, 2: 277-284. diabetics. Diabetes 1980, 29: 815-819. 79. Tooke, J.E., Lin, P.-E., Ostergren, J. et al. Effects of 58. Redisch, W., Rouen, L.R., Terry, E.N. et al. Microvascular intravenous insulin infusion on skin microcirculatory flow in changes in early diabetes mellitus. Adv Metab Disorders Type 1 diabetes. Int J Microcirc Clin Exp 1985, 4: 69-83. 1973, Suppl. 2: 383-390. 80. Tooke, J.E. & Rayman, G. The investigation of foot skin 59. Tubiana-Rufi, N., Priollet, P., Levy-Marchal, C. et al. microcirculation in diabetes mellitus. In: Spence, V.A. & Detection by nailfold capillary microscopy of early mor- Sheldon, C.D. (eds) Practical Aspects of Skin Blood Flow phologic capillary changes in children with insulin depen- Measurement. Conference Proceedings. Biological dent diabetes mellitus. Diabetes Metab 1985, 15: 118-122. Engineering Society, London, 1985, pp. 73-77. 60. Karlander, S.G., Hermansson, I.L. & Hellstrom, K. Nutri- 81. Rayman, G., Hassan, A. & Tooke, J.E. Blood flow in the tive toe skin capillaries in middle-aged patients with diabetes skin of the foot related to posture in diabetes mellitus. Br J mellitus. Diabetes Metab 1985, 2: 165-169. Med 1986, 292: 620-621 (reply to letter). 61. Kohner, E.M., Hamilton, A.M., Saunders, S.J. et al. The 82. Rayman, G., Spencer, P.D., Smaje, L.H. et al. Relationship retinal blood flow in diabetes. Diabetologia 1975,11: 27-33. ofcutaneous blood flow to duration and control ofdiabetes. 62. Felder, D., Russ, E., Montgomery, H. et al. Relationship in Int J Microcirc Clin Exp 1984, 3: 285 (Abstract). the toe of skin surface temperature to mean blood flow 83. Grunwald, J.E., Riva, C.E., Sinclair, S.H. et al. Laser measured with a plethysmograph. Clin Sci 1954, 13: Doppler velocimetry study of retinal circulation in diabetes 251-257. mellitus. Arch Ophthalmol 1986, 104: 991-996. 63. Wollersheim, H., Netten, P.M., Lutterman, J.A. et al. 84. Goodman, A.H., Guyton, A.C., Drake, R. et al. A television Ephedrine improves microcirculation in the diabetic neuro- method for measuring capillary red cell velocities. J Appl pathic foot. Angiology 1989, Dec: 1030-1034. Physiol 1974, 37: 126-130. 64. Nitzan, M., Goss, D.E., Chagne, D. et al. Assessment of 85. Intaglietta, M., Silverman, N.R. & Tompkins, W.R. Capil- regional blood flow and specific microvascular resistance in lary flow velocity measurements in vivo and in situ by by copyright. the foot by means ofthe transient thermal clearance method. television methods. Microvasc Res 1975, 10: 165-169. Clin Phys Physiol Meas 1988, 9: 347-352. 86. Chittenden, S.J., Coleridge Smith, P.D., Lefford, F. et al. An 65. Holti, G. & Mitchell, K.W. Estimation of the nutrient skin improved system for measuring capillary red cell velocities blood flow using a non-invasive segmented thermal clear- by video microscopy incorporating an automated vessel ance probe. In P. Rolfe (ed.) Non-invasive Physiological tracking facility. Int J Microcirc (Clin Exp) 1990,9: p. 135H Measurements, Vol. 1. Academic Press, London, 1979, (Abstract). pp. 113-123. 87. Chittenden, S.J. & Coleridge Smith, P.D. Measurement of 66. Scott, A.R., Bennett, T. & MacDonald, I.A. Diabetes blood cell movement in the capillaries. Binary Comp Micro- mellitus and thermoregulation. Can J Physiol Pharmacol biol 1991, 3: 107-110. 1987, 65: 1365-1376. 88. Tyml, K. & Sherebrin, M.H. A method for on-line measure- 67. C.W.I. & J.S. Measurement Corcoran, J.S., Owens, Yudkin, ment of red cell velocity in microvessels using computerized http://pmj.bmj.com/ of fingertip blood flow using thermal clearance reflects frame-by-frame analysis oftelevision images. Microvasc Res anastomotic rather than nutrient blood flow. Clin Sci 1987, 1980, 20: 1-8. 72: 225-232. 89. Tyml, K. & Ellis, C.G. Evaluation of the flying spot 68. Eickhoff, J.H. & Jacobsen, E. Correlation oftranscutaneous technique as a television method for measuring red cell oxygen tension to blood flow in heated skin. Scand J Clin velocity in microvessels. Int J Microcirc Clin Exp 1982, 1: Lab Invest 1980, 40: 761-765. 145-155. 69. Roberts, D.H., Tsao, Y. & Breckenridge, A.M. The repro- 90. Wayland, H. & Johnson, P.C. Erythrocyte velocity measure- ducibility of limb blood flow measurements in human ment in microvessels by a two-slit photometric method. J volunteers at rest and after exercise by using mercury-in- Appl Physiol 1967, 22: 333-337. Silastic strain gauge plethysmography under standardized 91. Fagrell, B., Fronek, A. & Intaglietta, M. A microscope- on September 28, 2021 by guest. Protected conditions. Clin Sci 1986, 70: 635-638. television system for studying flow velocity in human skin 70. Greenfield, A.D.M. A simple water-filled plethysmograph capillaries. Am J Physiol 1977, 233: H318-321. for the hand or forearm with temperature control. J Physiol 92. Katz, M.A. & Janjan, N. Forearm hemodynamics and (Lond) 1954, 123: 62-64. responses to exercise in middle-aged adult-onset diabetic 71. Whitney, R.J. The measurement of volume changes in patients. Diabetes 1978, 27: 726-731. human limbs. J Physiol (Lond) 1953, 121: 1-27. 93. Poulsen, H.L. & Nielsen, S.L. Water filtration ofthe forearm 72. Henriksen, O. Local sympathetic reflex mechanisms in in short- and long-term diabetes mellitus. Diabetologia 1976, regulation of blood flow in human subcutaneous adipose 12: 437-440. tissue. Acta Physiol Scand (Suppl) 1977, 450: 1-48. 94. Rayman, G., Hassan, A. & Tooke, J.E. Blood flow in the 73. Challoner, A.V.J. Photoelectric plethysmography for skin of the foot related to posture in diabetes mellitus. Br J estimating blood flow. In P. Rolfe (ed.) Non-invasive Med 1986, 292: 87-90. Physiological Measurements Vol. 1. Academic Press, 95. Cheatle, T.R., McMullin, G., Farrer, J. et al. Skin damage in London, 1979, pp. 36-60. chronic venous disease: does an oxygen diffusion barrier 74. Nilsson, G.E., Tenland, T. & Oberg, P.A. Evaluation of a really exist? J Roy Soc Med 1990, 83: 493-494. laser Doppler flowmeter for measurement of tissue blood 96. Breuer, H.-W.M., Breuer, J. & Berger, M. Transcutaneous flow. IEEE Trans Biomed Eng 1980, 27: 597-604. oxygen pressure measurements in type I diabetic patients for 75. Watkins, D. & Holloway, G.A., Jr. An instrument to early detection of functional microangiopathy. Eur J Clin measure cutaneous blood flow using the Doppler shift of Invest 1988, 18: 454-459. laser light. IEEE Trans Biomed Eng 1978, 25: 28-33. 428 S.J. CHITTENDEN & S.K. SHAMI Postgrad Med J: first published as 10.1136/pgmj.69.812.419 on 1 June 1993. Downloaded from

97. Huch, R., Lubbers, D.W. & Huch, A. Reliability of 103. Eickhoff, J., Ishihara, S. & Jacobsen, E. PaO2 by skin transcutaneous monitoring of arterial PO2 in newborn electrode. Lancet 1979, ii: 1188-1189. infants. Arch Dis Childhood 1974, 49: 213-218. 104. Falstie-Jensen, N., Spaun, E., Brochner-Mortensen, J. et al. 98. Mani, R., Gorman, F.W. & White, J.E. Transcutaneous The influence of epidermal thickness on transcutaneous measurements of oxygen tension at edges of leg ulcers: oxygen pressure measurements in normal persons. Scand J preliminary communication. J Roy Soc Med 1986, 79: Clin Lab Invest 1988, 48: 519-523. 650-654. 105. Fagrell, B. Microcirculatory disturbances - the final cause 99. McCollum, P.T., Spence, V.A. & Walker, W.F. Oxygen for venous leg ulcers? VASA 1982, 11: 101-103. inhalation induced changes in the skin as measured by 106. Franzeck, U., Bollinger, A., Huch, R. et al. Transcutaneous transcutaneous oxymetry. Br J Surg 1986, 73: 882-885. oxygen tension and capillary morphologic characteristics 100. Hauser, C.J. & Shoemaker, W.E. Use oftranscutaneous PO2 and density in patients with chronic venous incompetence. regional perfusion index to quantify tissue perfusion in Circulation 1984, 70: 806- 811. peripheral vascular disease. Ann Surg 1983, 197: 337-343. 107. Kobbah, A.M., Ewald, U. & Tuvemo, T. Impaired vascular 101. Eickhoff, J., Ishihara, S. & Jacobsen, E. Effect ofarterial and reactivity during the first two years ofdiabetes mellitus after venous pressures on transcutaneous oxygen tension. ScandJ initial resolution. Diabetes Res 1988, 8: 101-109. Clin Lab Invest 1980, 40: 755-760. 108. Dodd, H.J., Sarkany, I. & Gaylarde, P.M. Skin oxygen 102. Wyss, C.R,. Matsen, F.A. III, King, R.V. et al. Dependence tension in venous insufficiency of the lower leg. J Roy Soc of transcutaneous oxygen tension on local arteriovenous Med 1985, 78: 373-376. pressure gradient in normal subjects. Clin Sci 1981, 60: 499-506. by copyright. http://pmj.bmj.com/ on September 28, 2021 by guest. Protected