Volume 36, Issue 3 , Pages 346-352, September 2008
Significance of Phlebosclerosis in Non-healing Ischaemic Foot Ulcers of End-stage Renal Disease☆
Article Outline
Abstract
Objectives
To determine the influence of haemodialysis on the poor healing of ischaemic ulcers in end-stage renal failure patients regardless of successful revascularization.
Materials and methods
We investigated the microscopic findings of subcutaneous small vessels in the amputated limbs of 78 patients (27 diabetic/haemodialysis, 26 diabetic/non-haemodialysis and 25 non-diabetic/non-haemodialysis patients) who underwent foot/toe or limb amputation because of ischaemic foot ulcers in the period between 1998 and 2006. All the haemodialysis patients were diabetic. Multivariate logistic analysis was conducted to identify important clinical factors related to the histological findings.
Results
Marked medial thickening was observed in both small veins and arteries in diabetic patients compared with non-diabetic patients. In diabetics, there was significant medial thickening of small veins, which was greater in haemodialysis patients than in non-haemodialysis patients (Dunnett test, P
<0.05). Multivariate analysis indicated that haemodialysis treatment (odds ratio 14.12, P<0.01), ABI value (odds ratio 5.41, P<0.01) and poor stump wound healing (odds ratio 6.19, P=0.03) were important factors related to medial thickening of small veins.
Conclusions
Our data suggest that medial thickening of small veins, or phlebosclerosis, might affect the healing of ischaemic ulcers in end-stage renal failure, although the strong influence of diabetes cannot be ignored.
Keywords: End-stage renal disease, Haemodialysis, Non-healing ulcer, Diabetes, Phlebosclerosis
Introduction
In patients with end-stage renal failure the management of ischaemic ulcers is challenging. The ulcers are often persistent and slow to improve despite aggressive surgical treatment including distal pedal bypass or endovascular therapy.1, 2, 3
Several etiologic mechanisms have been proposed to explain why it is difficult to achieve healing of ulcers in end-stage renal disease. Haemodialysis patients are known to have severe arterial calcification. Several previous studies suggested that these calcified vessels are the main cause of non-healing ischaemic ulcers. However, luminal obstruction of small arteries in end-stage renal failure is not always associated with calcification.4
As most patients with haemodialysis are diabetic, the healing process of ischaemic ulcers should be affected by diabetic microangiopathy as well as atherosclerosis. Furthermore, diabetic ulcers are susceptible to infection, which also leads to the non-healing characteristics in end-stage renal failure. One study demonstrated a type of microangiopathy in the skin, so-called dermopathy.5 Another study supported the presence of a functional impairment of the microcirculation in diabetes.6 Stewart et al. demonstrated that functional impairment of endothelial cells leads to a decrease of skin perfusion in haemodialysis patients.7 Hinchliffe et al. reported that haemodialysis has a persistent influence on transcutaneous oxygen tension.8 However, the degree to which renal failure affects the healing process of diabetic ulcers in patients with end-stage renal failure remains unclear.
In terms of the microstructure of the skin, morphological alteration of the stroma has been demonstrated in haemodialysis patients. However, the precise mechanisms by which these changes lead to poor healing have not yet been delineated.
Thus, there has been little information about whether the non-healing ulcers of haemodialysis patients are due only to the influences of comorbid conditions such as diabetes and arteriosclerosis, or are related to haemodialysis itself. We studied the histological findings to examine the influence of haemodialysis on wound healing in end-stage renal failure.
Materials and Methods
Patients
We studied a total of 78 patients who underwent foot/toe or limb amputation because of ischaemic foot ulcers caused by diabetes or atherosclerosis at the Department of Surgery in Teikyo University Ichihara Hospital in the period between January 1998 and March 2006. The patient group consisted of 53 consecutive diabetic patients (27 haemodialysis patients and 26 non-haemodialysis patients) and 25 age- and sex-matched control patients. These patients included 59 men and 19 women with a mean age of 63.6 years (39–86 years). All the haemodialysis patients had type 2 diabetes and had received haemodialysis treatment three times a week in the dialysis centre of our hospital for a mean period of 6.0 years, ranging from 3–240 months.
The control group consisted of patients without diabetes and end-stage renal failure who underwent foot/toe or limb amputation because of arterial obstruction during the same study period. They served as age- and sex-matched controls to determine the influences of diabetes and haemodialysis.
Among the haemodialysis patients, 8 patients underwent vascular reconstruction. They included 7 cases of femoro-popliteal bypass (2 cases of below-knee bypass) and one of popliteal- pedal bypass. On the other hand, the 8 non-haemodialysis diabetic patients who underwent vascular reconstruction included 6 cases of femoro-popliteal bypass (one case of below-knee distal bypass) and 2 of femoro-popliteal bypass with iliac stenting. Furthermore, among the non-haemodialysis/non-diabetic patients, 12 patients underwent vascular reconstruction. Among those patients, femoro-popliteal bypass was the dominant procedure. A total of 7 patients underwent femoro-popliteal bypass. The other 5 patients included 3 cases of femoro-femoral cross-over bypass, one of aorto-femoral bypass, and one of aorto-femoral bypass with femoro-popliteal bypass.
Poor wound healing was defined as delayed removal of sutures more than 3 weeks after amputation because of any condition interfering with wound healing such as infection or insufficient blood flow.
ABI measurement was performed in a routine manner. As the perfusion pressure of the foot was usually low in the study patients, it was possible to measure ABI even in patients with arterial calcification. In several patients, we assessed ABI values following leg elevation.
Culture of swabs taken from ulcers and wounds was routinely performed at the initial presentation and several times during the course of treatment, even if concomitant infection was not suspected.
Tissue specimens
The histological samples consisted of 27 samples from the haemodialysis patients and 51 from the non-haemodialysis patients. All the samples were obtained from the adjacent skin proximal to the ischaemic foot/toe or limb, where the skin was not macroscopically affected by degenerative changes due to necrosis. Tissue specimens were blindly examined by two independent pathologists for thickening and calcification of subcutaneous arteries and veins smaller than the digital vessels, with a diameter of 40–70μm.
Assessment of histological findings
Each entire section was observed under a microscope at high magnification (×200) using hematoxylin & eosin and van Gieson's staining. Wall thickening of the arteries and veins was evaluated in five randomly selected fields. Evaluation was carried out for each layer of the vessels, i.e. tunica intima and tunica media. Thickening of the layer was scored according to four grades from 0 (none) to 3 (greatest). We regarded complete obstruction of a vessel due to thickening of the wall as grade 3. The final score for each layer was defined as the mean of the total scores in arteries or veins counted in five fields. Two pathologists reached agreement on the final evaluation, which they made for each tissue specimen in order to overcome selection bias.
Calcification of small vessels was also assessed using similar scoring with four grades from 0 to 3. In terms of the calcification of small vessels, the score for the whole wall, rather than for each layer, was obtained.
We assumed that the spatial distribution of subcutaneous small vessels is uniform as long as the samples are taken from vessels with similar diameters. When the samples included apparent ischaemic degenerative lesions microscopically, the site of counting was always selected in the intermediate zone between the viable and ischaemic part, in order to minimize the influence of ischaemic damage of the samples. With this manipulation, it was assumed that the score would reflect the degree of wall thickening directly affecting non-healing of the ulcer. The influence of proximal arterial obstruction on the morphological changes was analyzed statistically using ABI values.
Statistics
Data were analyzed with SAS version 9.1 (SAS Institute, Cary, NC). Univariate analysis was performed initially to evaluate the possible importance of clinical variables. The relevant variables with univariate P values less than 0.10 were selected for inclusion in the initial step of logistic analysis. Comparison was performed using Student's t-test or Mann-Whitney U test. Comparisons among proportions were made using Pearson's chi-squared statistic to identify variables possibly correlated with pathological findings. Fisher's exact test for 2×2 tables was used for small samples.
Logistic analysis was used to identify important variables relating to the histological findings among the background clinical data obtained from the hospital records. Those data included the cause of end-stage renal failure, age, sex, current smoking status, associated diabetes mellitus, associated hypertension, concurrent atrial fibrillation, previous history of ischaemic heart disease or cerebrovascular disease, and administered drugs. A previous history of ischaemic heart disease included previous myocardial infarction (as documented by electrocardiogram, enzymes, or history), angina, or positive findings of non-invasive studies (myocardial scintigraphy or echocardiography). A previous history of cerebrovascular disease included a history of abrupt onset of neurological deficit regardless of its duration. The ankle brachial index (ABI) before amputation was also included in the analysis. Data relating to age and duration of symptoms were classified into categorical variables.
Variables that were significantly associated with histological findings were entered into a logistic regression model, and the variables were assessed to determine which ones were independently associated with the histological findings, using stepwise manipulation. The level of significance was set at a P value less than 0.05.
Ethical guidelines
The human research review committee of Teikyo University Ichihara Hospital approved the protocol of the present study, and informed consent for permission to use histological samples was given by each patient.
Results
The patient demographics are presented in Table 1. All the groups were matched for age and sex. All of them had ischaemic ulcers that were proved to be contaminated by culture studies, and over 40% of the patients in each group had MRSA or Pseudomonas infection. However, there was no statistically significant difference in the frequencies of infection with these bacteria between the groups.
Table 1. Demographics of patients with limb/toe amputation because of peripheral arterial occlusive disease
| Variable | Haemodialysis patientsc (n=27) | Non-haemodialysis diabetic patients (n=26) | Non-haemodialysis/non-diabetic patients (n=25) | P value |
|---|---|---|---|---|
| Age (SD) | 61.3 (10.4) | 63.6 (9.3) | 64.0 (8.4) | 0.14 |
| Male sex | 20 (74.1%) | 17 (65.4%) | 22 (88.0%) | 0.17 |
| Duration of ulcerb | 3.0 | 2.5 | 2.6 | 0.90 |
| MRSA/Psedomonas infection | 11 (40.7%) | 11 (42.3%) | 13 (52.0%) | 0.68 |
| ABI (SD) | 0.25 (0.22) | 0.42 (0.42) | 0.15 (0.27) | 0.01 |
| Current smoking | 19 (70.4%) | 16 (61.5%) | 16 (64.0%) | 0.78 |
| Bypass surgery | 8 (29.6%) | 8 (30.8%) | 12 (48.0%) | 0.31 |
| Poor stump healing after amputation | 9 (33.3%) | 10 (38.5%) | 7 (28.0%) | 0.73 |
| Concurrent triopathya | 25 (92.6%) | 12 (46.2%) | 0 (0.0%) | <0.01 |
| Hypertension | 20 (74.1%) | 16 (61.5%) | 18 (72.0%) | 0.57 |
| History of ischaemic heart disease | 15 (55.6%) | 3 (11.5%) | 5 (20.0%) | <0.01 |
| Arrhythmia | 3 (11.1%) | 1 (3.9%) | 7 (28.0%) | 0.04 |
| History of cerebrovascular disease | 9 (33.3%) | 7 (26.9%) | 10 (40.0%) | 0.61 |
| Drug therapy | ||||
| ACE inhibitor | 8 (30.0%) | 5 (19.2%) | 4 (16.0%) | 0.46 |
| Ca2+ channel antagonist | 16 (59.3%) | 8 (30.8%) | 12 (48.0%) | 0.11 |
| Nitrates | 16 (59.3%) | 5 (19.2%) | 9 (36.0%) | 0.01 |
| Anti-platelet drug | 19 (70.4%) | 12 (46.2%) | 16 (64.0%) | 0.18 |
| Warfarin | 1 (3.7%) | 1 (3.9%) | 3 (12.0%) | 0.44 |
| Prostaglandin analogue | 26 (96.3%) | 22 (84.6%) | 24 (96.0%) | 0.31 |
| Digoxin | 1 (3.7%) | 1 (3.9%) | 4 (16.0%) | 0.27 |
aRetinopathy, neuropathy and nephropathy. |
bValues are presented as median. |
cAll haemodialysis patients were diabetic. |
The prevalence of a previous history of ischaemic heart disease was higher in haemodialysis patients than in non-haemodialysis patients. The amputations in non-haemodialysis patients consisted of 13 above-knee, 21 below-knee and 17 foot/toe amputations, while those in haemodialysis patients consisted of 7 above-knee, 11 below-knee, and 9 foot/toe amputations. The distribution of amputation was different between the diabetics and non-diabetics (Fisher exact test, P=.0012). There was a high frequency of major amputation above the ankle in non-diabetic compared with diabetic patients. However, the distribution of amputation did not differ significantly between the haemodialysis and non-haemodialysis patient groups (P=1.0000) (Table 2).
Table 2. Level of amputation in study patients
| Patient group | Haemodialysis patientsa (n=27) | Non-haemodialysis diabetic patients (n=26) | Non-haemodialysis/non-diabetic patients (n=25) |
|---|---|---|---|
| Above-knee amputation | 7 | 2 | 11 |
| Below-knee amputation | 11 | 9 | 12 |
| Foot/toe amputation | 9 | 15 | 2 |
aAll haemodialysis patients were diabetic. |
Results of histological assessment
Before assessment of the pathological results, the agreement of two pathologists was measured using kappa coefficient for each observation. The values were 0.87, 0.84, 0.82 and 0.84 for arterial intima, arterial media, venous intima, and venous media, respectively (data not shown). A high degree of agreement was shown between the two pathologists.
The mean score of medial thickening of the small arteries was significantly higher in the haemodialysis group than in the non-haemodialysis/non-diabetic group, while the scores of intimal wall thickening did not show a statistically significant difference. However, the score of medial thickening of the arteries in haemodialysis patients was similar to that in non-haemodialysis diabetic patients. Regarding small veins, the mean score of medial thickening in haemodialysis patients was significantly higher than that in both non-haemodialysis diabetics and non-haemodialysis/non-diabetic patients (Table 3) (Figure 1, Figure 2).
Table 3. Score of wall thickening in small arteries and veins
| Studied microvessels and their layers | Haemodialysis patientsd (n=27) | Non-haemodialysis diabetic patients (n=26) | Non-haemodialysis/non-diabetic patients (n=25) | P-valuea |
|---|---|---|---|---|
| Small arteries | ||||
| Intima | 0.69±0.68 | 0.62±0.64 | 0.48±0.60 | 0.51a |
| Media | 1.17±0.85b | 1.08±0.78 | 0.52±0.64b | <0.01a |
| Small veins | ||||
| Intima | 0.46±0.76 | 0.48±0.74 | 0.08±0.40 | 0.06a |
| Media | 1.24±1.05b, c | 0.77±0.90c | 0.28±0.56b | <0.01a |
aAnalysis of variance. |
bDunnett test, P |
cDunnett test, P |
dAll haemodialysis patients were diabetic. |
Figure 1
Typical photomicrographs of small vessels in non-haemodialysis, non-diabetic patients. Elastica von Gieson staining shows maintenance of three layers, i.e. intima, media, and adventitia, in an arteriole (left lower corner) and a small vein (right upper corner).
Figure 2
Typical photomicrographs of small veins showing medial thickening in haemodialysis patients (arrow).
There was no apparent calcification of small veins in either haemodialysis or diabetic patients, while calcification of small arteries was significantly greater in haemodialysis patients than in both non-haemodialysis diabetics and non-haemodialysis/non-diabetic patients (Mann-Whitney test, P<0.05).
Important clinical factors related to thickening of arterial wall
Multivariate analysis indicated that thickening of the arterial wall was related to associated diabetic triopathy (retinopathy, neuropathy and nephropathy), previous history of cerebrovascular disease and previous history of ischaemic heart disease (Table 4).
Table 4. Significant factors for medial thickening of small arteries by logistic regression analysis
| Variable | Df | Odds ratio (95% CI) | P value |
|---|---|---|---|
| Concurrent triopathya | 1 | 15.308 (2.539–92.303) | <0.01 |
| 0=no | |||
| 1=yes | |||
| History of cerebrovascular disease | 1 | 19.013 (3.581–100.952) | <0.01 |
| 0=no | |||
| 1=yes | |||
| History of ischaemic heart disease | 1 | 0.125 (0.020–0.792) | 0.02 |
| 0=no | |||
| 1=yes |
aRetinopathy, neuropathy and nephropathy. |
Important clinical factors related to thickening of venous wall
Multivariate analysis indicated that thickening of the venous wall was related to haemodialysis treatment, ABI value ≤0.3 and poor postoperative wound healing (Table 5).
Table 5. Significant factors for medial thickening of small veins by logistic regression analysis
| Variable | Df | Odds ratio (95% CI) | P value |
|---|---|---|---|
| Haemodialysis treatment | 1 | 14.12 (2.71–73.52) | <0.01 |
| 0=no | |||
| 1=yes | |||
| ABI value | 1 | 5.41 (1.12–26.11) | <0.01 |
| 0=above 0.3 | |||
| 1=equal or less than 0.3 | |||
| Poor wound healing | 1 | 6.19 (1.41–27.29) | 0.03 |
| 0=no | |||
| 1=yes |
Discussion
The healing process of ischaemic ulcers is impaired by various underlying conditions in diabetic patients. Those conditions include decreased blood perfusion in large and small vessels due to atherosclerosis, susceptibility to infection, and recurrent trauma due to diabetic neuropathy or the sequelae of cerebrovascular disease. Above all, diabetic patients with associated end-stage renal failure have poor wound healing in spite of successful bypass surgery.1, 2, 3 The most important finding of this study is that the mean score of medial thickening of small veins was significantly higher in haemodialysis patients than in non-haemodialysis patients.
Several histological studies have investigated the mechanisms of poor wound healing in diabetics. One previous study demonstrated marked PAS (periodic acid-Schiff stain) deposition in the basal membrane of capillaries.9 Another recent study showed increased apical and non-apical dilatation, hemorrhagic extravasations and branching of nail capillaries.10 The histopathological features of the diabetic foot include increased inflammatory elements, matrix alterations, vessel disruption, inflammation, and debris.11 In contrast to these morphological changes, many studies published over the last decade have revealed that functional rather than structural changes are more responsible for the poor wound healing, through impaired ability to vasodilate in response to injury.3, 12 The histological findings in the present study are not consistent with those previous reports, showing morphological changes in both small veins and arteries, but not in capillaries. These findings suggest the possibility that the histological changes in small veins involve distinct pathogenetic mechanisms from diabetic microangiopathy or so-called diabetic foot.
It is hard to determine whether our histological findings were caused by haemodialysis itself or by one of the diabetic sequelae, as all the haemodialysis patients were diabetic in the present study, although there is a tendency for thickening of small vessels to be more apparent in haemodialysis patients. To overcome this problem, we conducted multivariate analysis to confirm the independence of the haemodialysis factor in venous wall thickening. The results of logistic analysis indicated that haemodialysis was an independent factor for venous medial thickening (or phlebosclerosis), while the association of diabetic triopathy (retinopathy, neuropathy and nephropathy) and atherosclerotic complications were significant variables related to medial thickening of the small arteries. These results justify our conclusion that haemodialysis is an independent variable for venous sclerosis, and venous thickening appears to be related more closely to the poor wound healing in haemodialysis patients. Although the morphological changes in the small veins may be secondary, there is a possibility that non-healing ulcers in haemodialysis may be caused by venular congestion or venous hypertension,13 such as occurs in stasis ulcers.
The present study has a limitation in that our series was not matched for the level of amputation and comorbid conditions such as previous history of ischaemic heart disease, although all the patient groups were matched for age and sex. These patient characteristics may induce more advanced arteriosclerosis in haemodialysis patients. In fact, haemodialysis patients are well known to have greater sclerotic comorbidity. Although advanced arteriosclerosis might have influenced the outcome of the present study, atherosclerosis is not usually associated with phlebosclerosis. Therefore, it is not likely that artherosclerosis would have affected the development of fibrosis of the small veins.
There is a possibility that reduced perfusion of the foot due to proximal obstructive lesions may induce thickening of distal vessels. Anvar and colleagues14 examined by transmission electron microscopy skin biopsies taken from the ischaemic foot, and suggested that the stromal haemorrhage and degeneration caused by “gaps” between elongated oedematous endothelial cells probably signifies a terminal stage of critical limb ischemia. Another study by Coats15 pointed out functional inconsistencies between the proximal and distal arterioles in critical limb ischemia. These results are consistent with ours in the sense that perfusion status may influence the vessels of the foot not only functionally but also morphologically. However, neither of the studies specified the changes in small veins or venules, unlike the present study. Thus, we believe that the influence of proximal hemodynamics on distal vessels has not been delineated in previous studies. To overcome this problem, we used ABI measurement in statistical analysis. Although toe pressure reflects the perfusion status of the foot more precisely, we used ABI values because of the high frequency of necrotic changes of the toes. We attempted to minimize the bias in ABI measurement in non-compressible arteries by means of leg raising manipulation. Based on the results of the above mentioned analysis, we believe that the influence of the perfusion status of the foot is relatively small, if any, and a reduction of the perfusion alone cannot induce the microscopic changes in small veins observed in our study.
There are several possible causes leading to wall thickening of small veins.16 First, persistent infection may induce perivascular fibrosis surrounding small veins. Venous stasis ulcers may have a similar pathogenesis to non-healing ulcers in haemodialysis patients, although many previous studies have focused on morphological changes in the capillaries.17, 18 Secondarily, recurrent injury due to diabetic neuropathy might induce fibrosis of the venous wall. In fact, the prevalence of a previous history of cerebrovascular disease in haemodialysis patients was higher than that in non-haemodialysis patients in our series. Loss of sensation of the extremity due to cerebrovascular events may relate to these findings. Thirdly, there is a possibility that use of insulin may relate to thickening of the venous media. It is well known that insulin causes overgrowth and wall thickening of the retinal microvessels, acting as a growth hormone-like substance.19 Furthermore, insulin is reported to stimulate vasoconstriction.20, 21 Our microscopic findings could be consistent with those observed in diabetic retinopathy, although there are some reports that the thickness of the vessel wall in diabetes does not differ according to the use of insulin.22 The high incidence of retinopathy in patients with non-healing ulcers may support this speculation. Lastly, hypotension during haemodialysis treatment may induce vasoconstriction in the skin. Haemodialysis patients often experience recurrent hypotension during haemodialysis treatment, although it is not likely that hypotension affects the lower extremity alone. Haemodialysis patients also experience recurrent platelet activation by the filter as well as hemodynamic changes, which are additional influences of haemodialysis besides the uremic state. Although it is not easy to determine which of the above-mentioned factors are most important in the venous sclerosis observed in this particular study, the significant influences of diabetes can not be ruled out.
In conclusion, our data suggest that medial thickening of small veins, or phlebosclerosis might be one of the significant factors affecting the healing of ischaemic ulcers in end-stage renal failure, although the strong influence of diabetes cannot be ignored. Further study is needed to determine the main cause of venous wall thickening, with a non-diabetic haemodialysis group serving as a control.
References
- . Failure of foot salvage in patients with end-stage renal disease after surgical revascularization. J Vasc Surg. 1995;22:280–285
- . Outcome of infrainguinal bypass surgery for critical leg ischaemia in patients with chronic renal failure. Eur J Vasc Endovasc Surg. 1998;15:122–127
- . Cutaneous microcirculation in the neuropathic diabetic foot improves significantly but not completely after successful lower extremity revascularization. J Vasc Surg. 2002;35:501–505
- Atherosclerosis and vascular calcification are independent predictors of left ventricular hypertrophy in chronic haemodialysis patients. Nephrol Dial Transplant. 2005;20:760–767
- . Manifestations of cutaneous diabetic microangiopathy. Am J Clin Dermatol. 2005;6:225–237
- . Skin blood flow in diabetic dermopathy. Arch Dermatol. 2004;140:1248–1250
- Noninvasive interrogation of microvasculature for signs of endothelial dysfunction in patients with chronic renal failure. Am J Physiol Heart Circ Physiol. 2004;287:H2687–H2696
- . The effect of haemodialysis on transcutaneous oxygen tension in patients with diabetes – a pilot study. Nephrol Dial Transplant. 2006;21:1981–1983
- . Hypothesis on microangiopathy of cutaneous capillaries. Atherosclerosis. 1978;29:397–403
- . Vital capillary microscopic findings in the nailfold of patients with diabetes mellitus. Vasa. 2000;29:258–263
- Semiquantitative analysis of the histopathological features of the neuropathic foot ulcer: effects of pressure relief. Diabetes Care. 2003;26:3123–3128
- Endothelial dysfunction and the expression of endothelial nitric oxide synthetase in diabetic neuropathy, vascular disease, and foot ulceration. Diabetes. 1998;47:457–463
- . An animal model of venous hypertension: the role of inflammation in venous valve failure. J Vasc Surg. 2005;41:203–211
- . Vascular and stromal features in the skin of the lower limb in patients with chronic critical limb ischaemia (CLI) and Oedema. Eur J Vasc Endovasc Surg. 2000;20:125–131
- . Myogenic, mechanical and structural characteristics of resistance arterioles from healthy and ischaemic subjects. Clin Sci (Lond). 2003;105:683–689
- . Venous hypertension and the inflammatory cascade: major manifestations and trigger mechanisms. Angiology. 2005;56(Suppl. 1):S3–S10
- . Morphology of chronic venous insufficiency – light and electron microscopic examinations. Vasa. 1991;20:330–342
- . Microvascular changes in venous disease: an update. Angiology. 1997;48:27–32
- Acute intensive insulin therapy exacerbates diabetic blood-retinal barrier breakdown via hypoxia-inducible factor-1 alpha and VEGF. J Clin Invest. 2002;109:805–815
- . Modulation of vasoconstriction by insulin. J Hypertens. 1998;16:1157–1164
- . The role of insulin-like and basic fibroblast growth factors on ischaemic and infarcted myocardium: a mini review. Int J Cardiol. 1997;59:1–5
- . Quantitative evaluation of the dermal vasculature of diabetics. Q J Med. 1985;54:229–239
☆ Source of finance: None.
PII: S1078-5884(08)00269-4
doi:10.1016/j.ejvs.2008.05.003
© 2008 European Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.
Volume 36, Issue 3 , Pages 346-352, September 2008
