Comparison of pathological clotting using haematological, functional and morphological investigations in HIV-infected and HIV-uninfected patients with deep vein thrombosis

Background Patients infected with the human immunodeficiency virus (HIV) are more prone to systemic inflammation and pathological clotting, and many may develop deep vein thrombosis (DVT) as a result of this dysregulated inflammatory profile. Coagulation tests are not routinely performed unless there is a specific reason. Methods We recruited ten healthy control subjects, 35 HIV negative patients with deep vein thrombosis (HIV negative-DVT), and 13 HIV patients with DVT (HIV positive-DVT) on the primary antiretroviral therapy (ARV) regimen- Emtricitabine, Tenofovir and Efavirenz. Serum inflammatory markers, haematological results, viscoelastic properties (using thromboelastography-TEG) and scanning electron microscopy (SEM) of whole blood (WB) were used to compare the groups. Results DVT patients (HIV positive and HIV negative) have raised inflammatory markers. The HIV positive-DVT group has anaemia in keeping with anaemia of chronic disorders. DVT patients have a hypercoagulable profile on the TEG but no significant difference between HIV negative-DVT and HIV positive-DVT groups. The TEG analysis compared well and supported our ultrastructural results. Scanning electron microscopy of DVT patient’s red blood cells (RBCs) and platelets demonstrates inflammatory changes including abnormal cell shapes, irregular membranes and microparticle formation. All the ultrastructural changes were more prominent in the HIV positive-DVT patients. Conclusions It is well-known that HIV infection is linked to inflammation and inflammation is linked with the presence of a hypercoagulable state. The presence of DVT is also associated with inflammation. Whether HIV is the cause of the DVT is not certain. Although there were trends that HIV infected patients were more hypercoagulable on functional tests (viscoelastic profile) compared to HIV uninfected patients, there were no significant differences between the 2 groups. Morphologically there were inflammatory changes in patients with DVT. These ultrastructural changes,

During HIV infection, various circulating inflammatory biomarkers, including cytokines interleukin (IL)-1β, IL-2, IL-6, IL-8, IL-10, IL-12p70, tumor necrosis factor (TNF)-α and also other pro-inflammatory biomarkers are present (9). An increase in these biomarkers are also present in cardiovascular disease (10,11) and it is therefore not surprizing that HIV positive individuals are known to have an increased presence of cardiovascular complications (12,13), including an increased risk to develop atherosclerosis and venous thromboembolic disease (14) and also DVT (15)(16)(17). The presence of DVT is also classified as a systemic inflammatory process (18), and associated with pathological clotting and upregulated circulating inflammatory biomarkers (19).
The prevalence of developing a DVT in HIV positive individuals is increased two to ten times compared to the general population (20). HIV positive individuals also have a 43% increase in age-adjusted odds ratio for pulmonary embolism, a common complication of DVT, compared to HIV negative individuals (21). Multiple coagulation abnormalities have been reported in HIV positive patients such as decreased levels of protein C and S; and increased levels of von Willebrand factor (22)(23)(24). However, the association of these abnormalities with DVT is not always consistent (22,25). Coagulation investigations are therefore not performed routinely in patients with HIV infection. Standard coagulation investigations are also not performed routinely as part of the management in patients with DVT, with the exception of a D-dimer which is used to assist with the diagnosis (26)(27)(28)).
In the current paper, we therefore study the haematological profiles, including clotting and various inflammatory markers, in the presence of DVT in HIV positive and HIV negative individuals and compare the results to that of healthy individuals. We compare inflammatory markers for iron (iron saturation, transferrin and serum ferritin), fibrinogen, high-sensitive C-reactive protein (CRP), erythrocyte sedimentation rate (ESR) and haematology analyser results, together with viscoelastic properties of whole blood (WB) and platelet poor plasma (PPP). We also looked at ultrastructure of platelets and erythrocytes/red blood cells (RBCs) (using whole blood smears) with the SEM, as well as after thrombin was added to whole blood, to study clot structure.

Materials And Methods
The aim was to compare the inflammatory and haematological profile of HIV patients with 5 DVT to HIV negative patients with DVT. An analytical and descriptive prospective case control study was used from 2 Pretoria academic hospitals, Kalafong Provincial Tertiary and Steve Biko Academic Hospital. Ten healthy control subjects, 35 HIV negative patients with DVT (HIV negative-DVT), and 13 HIV patients with DVT (HIV positive-DVT) on the primary ARV regimen-Emtricitabine, Tenofovir and Efavirenz-were recruited for the study. For each individual, five blood tubes of venous blood were drawn (this included ethylenediamine tetraacetic acid, buffered tri-sodium citrate and serum separator tubes).

Inflammatory marker analysis
Serum iron (total iron in blood) levels were measured together with iron saturation, transferrin (iron binding protein) and serum ferritin (iron storage form). Serum iron levels were measured by a modification of the automated AAII-25 colorimetric method.
Fibrinogen (quantitative measurement of functional fibrinogen by automated coagulation analysers), CRP (measured by latex-enhanced nephelometry) and ESR (measured by an automated ESR analyser) levels were also assessed.
Haematology analysis 6 A haematology analyser (Advia 2120i, Siemens Healthcare) was used to do full blood counts, and the analysis included white cell count (and its differential count), RBC count, haemoglobin level, haematocrit, mean corpuscular volume (MCV), mean corpuscular haemoglobin (MCH), the mean corpuscular haemoglobin concentration (MCHC), as well as platelet count and mean platelet volume (MPV).
Viscoelastic properties of whole blood and platelet poor plasma using thromboelastography Citrated WB, as well as PPP were used. Whole blood, collected in a citrate tube, was centrifuged to obtain PPP (15 minutes at 3000g). Whole blood was used to assess the full coagulation process, while PPP was used to assess coagulation without the influence of platelets and RBCs on the viscoelastic properties of the clot. Calcium chloride was added to either WB or PPP and 7 different parameters measured, which included: reaction time

Ultrastructure of platelets and red blood cells (RBCs)
The ultrastructure of platelets and RBCs were studied after preparing whole blood smears for scanning electron microscopy (SEM). 10 µl of WB was placed directly on a glass microscope slide, followed by fixing in 2.5% glutaraldehyde, dehydration (as per usual SEM preparation) (29) and mounting. Micrographs were taken with Zeiss Crossbeam 540 Field Emission Gun Scanning Electron Microscopy.

Statistical analysis
Graphpad 5 was used to do one-way ANOVA analysis. A post-test to compare groups was performed using Tukey's multiple comparison test.

Results
7 Table 1 demonstrates the demographics of the study. Inflammatory marker analyses are shown in Table 2 and haematology analysis are shown in Table 3 Inflammation is reflected, whether from the DVT or the HIV infection, by the raised CRP 8 and ESR. Surprisingly, the platelet count was not decreased in the HIV positive-DVT group. We expected this parameter, as well as the MPV to be markedly decreased, due to e.g. HIV thrombocytopaenia, which is usually common amongst HIV patients, but in our sample this was not the case.  Thromboelastography Table 4 shows a comparison of the WB and PPP TEG results between the various groups. The

Discussion
Both DVT groups (HIV negative and HIV positive groups) had parameters suggesting anaemia compared to the control group (Table 3). However, the RBC count, Hb and Hct levels in the HIV negative-DVT group, even though lower than the control group, are still within the normal reference ranges (30). The Hb and Hct in the HIV positive-DVT group were significantly lower than both the HIV negative-DVT and control groups; and lower than the normal reference ranges indicating an anaemia. Anaemia is commonly found in 14 HIV positive patients but the cause of the anaemia is not always clear (31)(32)(33)(34)(35)(36)). An inadequate erythropoietin feedback mechanism is suspected to be a major contributor in HIV-related anaemia (31). A low reticulocyte count is commonly found with associated polychromasia (abnormally high number of immature RBCs), indicating a possible underproducing bone marrow (31,(37)(38)(39). Other factors that contribute to HIV-associated anaemia, includes intestinal malabsorption, autoimmune haemolysis, bone marrow malignancies, blood loss and opportunistic complications (31,33,39,40). Even with the decreased RBC count, Hb and Hct levels in the HIV negative-DVT group (as compared to the control group), there were no significant differences with MCV, MCH as well as MCHC ( Table 3). The changes in the HIV negative-DVT group may support an anaemia typically associated with inflammation, also known as anaemia of chronic disorders (31,41).
The RCDW is the coefficient of variation of RBC volume. The higher the value, the more anisocytosis (unequal RBC sizes) present. The RCDW of the HIV positive-DVT group was greater than the control and HIV negative-DVT groups ( Table 3). A raised RCDW is commonly associated with a decrease in haemoglobin and MCV concentration; but with a raised CRP, fibrinogen and white cell count (42). This correlates with the haematological and inflammatory markers found in the HIV positive-DVT group ( Table 2 and 3). RCDW is strongly associated with mortality. Patel and colleagues reported the all-cause mortality risk increases by 22% for every 1% increase in RCDW. Furthermore, the physiological association between RCDW and mortality has been hypothesised to be related to the systemic factors involved in inflammatory conditions and oxidative stress which affects erythrocyte maturation and degradation (42)(43)(44).
The ESR is the extent in which erythrocytes sediment in one hour (45). The ESR in both the HIV negative-DVT and HIV positive-DVT groups were raised compared to the control group ( Table 2). In inflammatory conditions the ESR rises as the erythrocytes become sticky and adhere to each other which can be seen as rouleaux formation (46)(47)(48).
Fibrinogen, a high molecular weight plasma protein, is a crucial factor in the coagulation pathway (factor I). Increased fibrinogen levels are associated with thrombo-embolic events. Fibrinogen also has a role in inflammation as it tends to adhere to the membrane receptors of cells involved with inflammation. Fibrinogen can adhere to the RBCs, which becomes "heavier" resulting in an increased ESR and blood viscosity (49)(50)(51)(52)(53)(54)(55)(56). The fibrinogen levels were greater (but not statistically significant) in the HIV negative-DVT and HIV positive-DVT groups compared to the control group ( Table 2)

16
The HIV negative-DVT group has an inflammatory response to the DVT which is reflected by the statistically significantly raised CRP levels compared to the control group ( Table   2). The same argument can be made for the raised CRP in the HIV positive-DVT group, however the CRP concentration (as well as fibrinogen) is commonly raised in HIV positive patients compared to the general population even without a DVT (65)(66)(67)(68)(69). The raised CRP in HIV positive-DVT patients ( Table 2) indicates a sustained acute phase response (67).
This was statistically significant in the HIV positive-DVT group compared to the control group. The CRP in the HIV positive-DVT group was almost double compared to the HIV negative-DVT group. Increasing CRP concentrations has been reported with HIV disease progression, and this increase does not appear to be affected by ARV treatment (64).
Previously it was noted that increased levels of CRP and fibrinogen are independently highly predictive of 5 year mortality risk in HIV positive patients, especially where the CD4 count is low (70,71).
Considering all the inflammatory markers (WCC, fibrinogen, CRP and ESR), each marker was statistically significantly raised in the HIV positive-DVT group compared to the control group, with the exception of fibrinogen ( Table 2). In the HIV negative-DVT group compared to the control group, only CRP was statistically significantly raised. CRP may therefore be a more sensitive acute phase marker to differentiate an inflammatory condition between DVT patients (HIV negative and HIV positive) compared to healthy subjects. Interestingly, no inflammatory marker was statistically significantly raised in the HIV positive-DVT group compared to the HIV negative-DVT group.
The transferrin, serum iron and iron saturation levels reflects the amount of iron in the body. Transferrin is a plasma protein that transports iron in the blood (63), whereas ferritin is an intracellular structure capable of storing iron atoms. The concentration of serum ferritin is related to the reticuloendothelial iron stores (72). Serum ferritin and iron concentrations are also indicators for acute phase responses to inflammation (72), although serum ferritin appears to be a better marker of inflammation than iron status (73).
Iron deficiency may be a contributor to anaemia in the HIV positive-DVT group in keeping with a low MCH and MCHC, although this is not reflected with the MCV which was within the normal reference range (Table 3) (74,75). A low serum iron and transferrin level seen in the HIV positive-DVT group, but with a raised ferritin level (as compared to the control group), can be explained by an immunologically altered iron metabolism where the body has adequate or increased iron stores but is unable to utilize those stores (33,38,61,62,72,(76)(77)(78). This functional iron deficiency can be considered a host defence mechanism by withholding iron from possible pathogens (79,80). However, as iron is required for normal immune function, iron deficiency can also increase the risk of infection (80).
Although the inflammatory RBC changes have been documented in non-communicable diseases, there are only a few reports of communicable diseases, specifically HIV, and the effect on RBCs and the coagulation system (81)(82)(83)(84)(85)(86). Multiple abnormal RBC shape changes and membrane abnormalities were noted in the patients with DVT (HIV negative and HIV positive groups) (Figure 1 to 3). During inflammatory diseases, RBCs exposed to oxidative stress and inflammatory molecules undergoes biochemical membrane changes which can result in biophysical shape changes and eryptotic cells (87)(88)(89)(90)(91)(92)(93)(94)(95). Eryptosis is a co-ordinated suicidal death of the red blood cells, similar to apoptosis, that allows for the removal of defective, infected or potentially harmful cells before they undergo haemolysis (96)(97)(98)(99)(100). The abnormal RBCs present with an abnormal expression of phosphatidylserine, a cell membrane lipid, on the external membrane layer. RBCs that display phosphatidylserine also contribute to the hypercoagulation state and they provide a prothrombotic surface for the formation of thrombin (41,98,(101)(102)(103)(104)(105)(106)(107)(108)(109)(110)(111)(112). Membrane vesicle formation and microparticle shedding (microscopic extracellular membranous structures) were also seen in both DVT groups. RBC-derived microvesicles or microparticles, is known to be associated with the expression of phosphatidylserine (113). RBC-derived microparticles appear to enhance thrombin generation resulting in a hypercoagulable state, such as in post transfusion DVT, sickle cell disease and haemolytic anaemia (114,115). As the microparticle presence might also be associated with increased thrombin presence, the complement system can therefore also be activated and thereby enhance the systemic inflammatory response which is also a hypercoagulable state (116).
Microparticles are also thought to originate from CD4 lymphocytes (117). As the HIV virus infects CD4 lymphocytes, HIV positive patients may be more prone to developing microparticles and therefore enhancing the hypercoagulable state.
Whole blood with thrombin SEM analysis showed the incorporation of RBCs into the fibrin network. The incorporation of RBCs into the fibrin network stabilises and strengthens the clot by decreasing the permeability of fibrin and increasing the resistance to fibrinolysis (118)(119)(120). Healthy (discoid) RBCs in netted fibrin fibers are shown Figure 1C and 1D.
However, in our HIV negative-DVT and HIV positive-DVT groups, the RBCs are trapped in a matted fibrin fiber network. During inflammation, fibrin fibres tend to increase in diameter and assume a matted rather than a netted appearance; while their viscoelasticity may also be influenced by the RBC inclusion in the fibrin network (101). Also, under conditions of low partial pressure of oxygen, acidosis and in response to mechanical deformation, RBCs release ATP and ADP activating platelets and promoting aggregation and release of platelet granules (115). This can happen as part of the HIV and DVT pathology. The (hyper) activation of platelets, together with an abnormal matted fibrin network, contracts the clot containing the trapped pathological RBCs into a tight package (Figure 3B and  3C). The result is the formation of polyhedrocytes, which is commonly found in DVT (121).
Platelet functioning depends on the quality and the quantity of the platelets (122).
Platelet count is a measure of the number of platelets in a volume of blood.
Thrombocytopenia (low platelet count) is a common finding in HIV positive patients, whether it be due to increased destruction or decrease production of platelet cells (57).
However, in this study both the HIV negative-DVT and HIV positive-DVT groups had a nonstatistically significant increase in the platelet count ( Table 3). It should be noted that platelet count is not always associated with an increased risk of DVT (123). The mean platelet volume measures the average size of platelets in the blood and is a common platelet activation marker (123)(124)(125)(126)(127)(128)(129)(130). An elevated MPV is associated with low-grade inflammation as well as thrombosis (131). However, both HIV negative-DVT and HIV positive-DVT groups had a decrease in the mean platelet volume compared to the control group ( Table 3). These results may be in keeping with a venous thrombosis where the thrombus is due to activation of the coagulation cascade instead of platelets (132). It should also be kept in mind that platelets shape and volume do vary, resulting in changes in MPV, even in healthy persons (130). Together with these results, the ultrastructure of platelets in the HIV positive-DVT group also have features different to that of the control group and the HIV negative-DVT group (Figure 1B, 2D and 3D Our TEG analysis compared well and supported our ultrastructural results (

Availability of data and materials
The dataset(s) supporting the conclusions of this article are available from the authors.

Competing interests
The authors declare no competing interests.