The Role of Inflammation in Venous Disease

PD Coleridge Smith FRCS, CM Butler RGN, JH Scurr FRCS,

Department of Surgery, University College London Medical School, The Middlesex Hospital, Mortimer Street, London W1N 8A

Contents

Abstract
Venous valvular incompetence is present in up to 10% of the population of Western countries and 0.2% have venous ulceration as a consequence. It is generally agreed that sustained venous hypertension is an essential pre-requisite to the development of ulcers but the precise patho-physiological mechanism remains obscure. Various theories of the aetiology of venous ulceration have been put forward, of which the two currently most popular are the fibrin cuff theory of Browse and Burnand and the white cell trapping theory of Coleridge Smith et al. Research has failed to show that fibrin cuffs present a diffusion barrier to the passage of oxygen "and other nutrients", and cutaneous oxygen tension is actually higher in patients with venous disease. The white cell trapping hypothesis suggests that increased venous pressure causes a reduction in capillary blood flow, resulting in trapping of the white blood cells in the leg. These cells may plug the capillaries, resulting in areas of localised ischaemia, and become activated, releasing toxic oxygen metabolites (free radicals), proteolytic enzymes and chemotactic substances. White cell margination has been shown to occur in the post-capillary venules when blood flow is reduced, and increased numbers of white cells have been demonstrated in the skin of patients with lipodermatosclerosis (LDS) We have recently obtained direct evidence for the release of products of white cell activation in venous disease. An understanding of the mechanisms responsible for venous ulceration may lead to the development of improved methods of treatment.

The "White Cell Trapping" hypothesis.
The search for the mechanisms of skin damage in venous disease has resulted in investigation of the blood. Moyses et al studied the limbs of normal subjects in response to raised venous pressure, and measured haematological parameters as an index of the effect of the venous hypertension . Their subjects sat on a bicycle saddle with the limbs dependent for a period of forty minutes without moving. Blood samples were taken from the long saphenous vein at the ankle. He found that the haematocrit and red cell count increased in parallel. However, the white cell count remained unchanged, despite the increased haematocrit. White cells were being 'lost' from the circulation, which after 40 minutes accounted for a 25% change in the white cell count. Thomas performed a similar study in which he compared patients with normal lower limbs to those of patients with venous disease resulting in lipodermatosclerosis and ulceration . Their patients sat with their legs dependent, a less stringent requirement than that of Moyses et al. Blood samples were taken from the long saphenous vein at the ankle. After 60 minutes patients with venous disease were 'trapping' 30% of the white cells and control subjects were trapping 7%. The white cells were "released" when patients lay supine. This lead us to examine the microcirculation using capillary microscopy. The number of visible capillary loops per unit area was determined in patients with normal limbs, varicose veins, and lipodermatosclerosis. Only in the patients with lipodermatosclerosis were there significant alterations in capillary counts, where there appeared to be fewer capillary loops. The number of loops decreased with increased venous pressure. We suspected that this was a consequence of the white cell trapping showed by Thomas et al. We found that venous hypertension appeared to reduce the number of visible capillary loops in patients with venous disease, but not in control subjects , suggesting that capillary damage may occur during venous hypertension.

We published a hypothesis suggesting that white cell trapping resulted in such processes being triggered, causing degradation of tissues (fig 1). In keeping with the literature on myocardium in critical ischaemia, we proposed that white cells might cause occlusion of capillaries, a suggestion originally made by Moyses et al1 . If some of the capillaries were occluded this might result in heterogeneous perfusion and therefore tissue hypoxia and ischaemia

Bollinger et al have investigated the events in venous disease using fluorescence video capillary microscopy . They measured the rate of diffusion of fluorescein out of capillaries after an intravenous injection. They showed that capillaries in venous disease are much more permeable than normal to this molecule, contrary to the suggestions made in the fibrin cuff hypothesis. Using simultaneous fluorescence and light capillary microscopy Franzeck et al have described the appearances of capillary loops which are filled with red blood cells, but do not appear to be perfused . They suggest that this may be due to capillary 'thrombosis'. After investigation by xenon clearance and oxygen return times , , we have not been able to support earlier hypotheses suggesting that venous ulceration is due to tissue hypoxia , .

The second part of the theory proposed that white cell activation was part of the process, resulting in release of proteolytic enzymes, superoxide radicals and chemotactic substances. All classes of white cells appear to become trapped so a wide range of phenomena is possible. Monocytes might become activated releasing the cytokines interleukin 1 (IL-1) and tumour necrosis factor alpha (TNF*) . These may cause endothelial cell activation, in which the endothelium permits the passage of much larger molecules than would normally be the case . Burnand discovered that the capillaries of patients with venous disease are much more permeable to fibrinogen than normal . In addition, he also noticed that fibrinolysis is decreased in patients with venous disease . IL-1 acts on endothelial cells to stimulate production of the fibrinolytic inhibitor plasminogen activator inhibitor 1 (PAI-1), and decreases the production of tissue plasminogen activator (tPA) producing the effect on fibrinolysis that is observed.

White cell margination is a normal event in the arterioles, capillaries and venules. In reperfusion injury it is thought to be important in the mechanism that results in tissue injury following ischaemia. White blood cells are substantially larger than red cells and are responsible for many of the rheological properties of blood. White cells take 1000 times longer than red cells to deform on entering a capillary bed, and are responsible for about half the peripheral vascular resistance despite their small numbers in the circulation compared with red cells . In myocardial infarction it has been shown that they cause capillary occlusion, which can be prevented in experimental animals by first rendering the animal leucopoenic , . White blood cells have been implicated as the mediators of ischaemia in many tissues including myocardium, brain, lung and kidneys - . Polymorphonuclear leukocytes, particularly those attached to capillary endothelium, may become 'activated' in which cytoplasmic granules containing proteolytic enzymes are released . In addition a non-mitochondrial 'respiratory burst' permits these cells to release free radicals, most notably, the superoxide radical, which have non-specific destructive effects on lipid membranes, proteins and many connective tissue compounds . The chemotactic leukotrienes are also released attracting more polymorphonuclear cells.

Histological Studies
Dermatology texts on venous disease describe infiltration of the skin by inflammatory cells. In order to investigate this quantitatively we took biopsies of the skin of the supra-malleolar region of patients undergoing varicose vein surgery. Three groups of patients were studied. The first were patients with no evidence of skin changes as a consequence of their venous disease. The next group exhibited lipodermatosclerosis, but there had never been ulceration of the limb. Finally, there was a group of patients who had had ulceration, but were left with lipodermatosclerosis after healing of the ulcer .

Skin biopsies were taken from the liposclerotic area and histologic slides made. The number of white blood cells visible in the upper 0.5mm of the skin in each section was estimated by an observer who was unaware of the diagnosis. The results are shown in Figure 2. Patients with normal skin had a low number of white blood cells visible (4 /sq. mm) . There were eight times as many in patients with liposclerotic skin and 40 times as many in patients with healed venous ulcers. We have subsequently undertaken an immunohistological study to determine the types of white cell present in this infiltrate. The majority of cells are macrophages with a T-lymphocyte component, but no excess of neutrophils compared with control sections taken from normal limbs. This infiltrate is a reflection of a chronic inflammatory process, and suggests that an investigation of the cell products of these leukocytes might indicate the mechanisms involved in venous ulceration. We have also been able to identify IL-1 as an inflammatory mediator in this process using immunohistochemical methods .

White blood cell metabolism in venous disease.
Histological studies demonstrate the presence of white cells, but do not indicate whether they are activated. Recently McCollum has demonstrated upregulation of neutrophil production of free oxygen radicals and increased production of thromboxane A2 in response to raised venous pressure in blood taken from the leg veins of patients with chronic venous insufficiency (McCollum CN)

Neutrophil elastase as an indicator of neutrophil activation.
Methods of assessing neutrophil function in the peripheral blood have been described that depend on the release of proteinases into the blood by activated leucocytes - . During the process of phagocytosis or when activated by other stimuli such as soluble immune complexes, C5a or endotoxins, neutrophils release several different lysosomal proteinases, of which neutrophil elastase appears to be the principal component , . The level of neutrophil elastase in the peripheral blood reflects neutrophil activity anywhere in the body and, therefore, provides a reliable, repeatable and relatively non-invasive measure of neutrophil activation.

We measured plasma elastase as a marker of neutrophil degranulation in three groups of 15 patients with uncomplicated varicose veins, lipodermatosclerosis (LDS) and venous ulceration and compared the values obtained with those in age- and sex-matched control subjects. Blood was taken from an arm vein in all patients and control subjects for full blood count including neutrophil count, erythrocyte sedimentation rate and plasma neutrophil elastase, measured using a radio-immunoassay. Higher levels of elastase were found in all patient groups compared to their controls (median 25.6 ng/ml. for uncomplicated varicose veins, controls 18 ng/ml., 22.1 ng/ml. for patients with lipodermatosclerosis, controls 17.7 ng/ml., and 26 ng/ml. for patients with venous ulceration, controls 18.8 ng/ml.), reaching statistical significance for all three groups (fig 3). There was no difference in the neutrophil count between the patient and control groups. This provides evidence of increased neutrophil degranulation in patients with venous disease, and the finding of raised levels in all three patient groups shows that this was not solely due to the inflammatory process found in LDS and venous ulceration. It is likely that the cause for this is neutrophil activation within the lower limb caused by venous hypertension.

Plasma Lactoferrin levels as indicators of neutrophil activation.
Plasma lactoferrin levels have been used as an indicator of neutrophil activation in several disease states including dental pulp disease , various forms of arthritis , and sepsis . This glycoprotein is found in several glandular epithelial tissues and human neutrophils, where it is localised to secondary granules. We measured plasma lactoferrin as a marker of neutrophil degranulation in groups of 10 patients with varying severity of venous disease and age- and sex-matched control subjects. We investigated 4 groups of patients with varicose veins, liposclerotic skin changes, active ulceration and healed ulcers compared to groups of control subjects with no history or clinical findings of venous disease.

Blood was taken from an arm vein for neutrophil count and plasma lactoferrin, measured using an enzyme-linked immunosorbant assay (ELISA). We found significantly raised plasma lactoferrin in all four groups of patients compared to their controls (p=.0156 for uncomplicated varicose veins, p=.01 for lipodermatosclerosis, p=.0413 for active venous ulceration, and p=.0005 for healed ulcers, Mann-Whitney U Test. Difference between medians (95% confidence interval) for the four groups were 269(62-603), 199(60-314), 133(44-218) and 215(98-349) ng/ml respectively). There was no difference in the neutrophil count between the patient and control groups. This study shows further evidence of increased neutrophil activation demonstrated by increased neutrophil degranulation in patients with venous disease.

Does short-term venous hypertension cause neutrophil activation?
Venous hypertension should cause neutrophil activation within one hour if the "white cell trapping" hypotheses is tenable. We measured lactoferrin levels in a group of 15 healthy normal volunteers with no clinical evidence of venous disease using two models of venous hypertension. A dorsal foot vein or lower long saphenous vein of both feet were cannulated using an 18G cannula (Venflon, Viggo-Spectramed, Helsingborg, Sweden), together with a vein on the dorsum of the right hand. The cannulae were then flushed with heparinised saline. Volunteers rested supine for 30 minutes on a couch. Two ml of blood were taken from each cannula and discarded, and a further 10 ml blood taken into two EDTA tubes for analysis. A wide pneumatic tourniquet was applied around the right thigh and inflated to 80 mmHg for 30 minutes. Further blood samples were taken from the three cannulae as above, the tourniquet deflated, and five minutes after deflation three further samples taken. The volunteers then rested for a further 30 minute period lying supine on the couch, following which they stood up for 30 minutes, resting against the edge of the couch such that no movement of the calf muscle pump was required to stay upright. At the end of this period a fifth set of blood samples were taken, and the experiment ended. Direct measurement of venous pressure showed that this protocol resulted in a pressure of 70-75 mmHg in the foot veins, without reducing arterial pressure. An ELISA developed by Dr JB Porter in the Department of Haematology, UCMSM, London was then used to assess lactoferrin levels. There was a significant rise in lactoferrin in the right leg after application of a tourniquet for 30 minutes but not in the left leg or arm (fig 4a).

During the second part of the experiment (standing for 30 minutes), lactoferrin levels increased significantly in all three limbs. The data for both legs added together show a rise in plasma lactoferrin (difference between medians 5.73 ng/ml, 95% confidence interval 1.2 to 10 ng/ml) (fig. 4b). These data confirm that short term venous hypertension results in neutrophil degranulation, in subjects with no evidence of venous disease.

We propose that in patients with venous valvular damage, repeated exposure of the lower limb to neutrophil activation may initiate the trophic skin changes seen in chromic venous insufficiency.

Conclusions
The precise mechanisms through which venous hypertension causes ulceration still remain unclear. Our investigation of white blood cell metabolism leads us to suspect that neutrophil activation may play an important role in initiating skin damage in patients with chronic venous insufficiency. A better understanding of the processes initiating this problem may lead to improvements in the management of patients with venous ulceration.

Acknowledgements
Our grateful thanks to the following members of the Department of Surgery of University College London Medical School who have undertaken much of the work described in this paper and made substantial intellectual contribution to our understanding of venous disease: Mr TR Cheatle, Mr HJ Scott, Mr DA Shields and Mr A Andaz.

Figure legends

Fig. 1 The 'White cell trapping' hypotheses summarised diagramatically.

Fig. 2. Results of quantitative histological study published by Scott 24. The vertical axis shows the number of white blood cells counted per square millimetre of histological section in the upper 1mm of the dermis.

Fig. 3 Results of measurement of plasma neutrophil elastase levels in patients with venous disease and control subjects (fig 3a). In fig 3b patients and age matched controls have been separated into those with uncomplicated varicose veins (VVs), lipodermatosclerosis (LDS) and active ulceration (Ulcer). Descriptors are medians and inter-quartile ranges.

Fig 4. Results of measurement of plasma neutrophil lactoferrin levels in normal volunteers subjected to experiment venous hypertension. Results are expressed at lactoferrin corrected for neutrophil count (ng/ml/109 neutrophils/l). Fig 4a shows the effect of 30 minutes of venous hypertension applied to the right lower limb using a cuff. Fig 4b shows the effect of standing for 30 minutes to produce venous hypertension. Numerical data is the median difference between data groups and the 95% confidence interval (95% CI).

References

Contents