The effects of plasma viscosity in thromboembolic events among patients with essential thrombocytosis: A case-control study

INTRODUCTION: Essential thrombocythemia (ET) is an entity of classic Philadelphia chromosome-negative myeloproliferative neoplasms (MPNs), characterized by thrombocytosis with megakaryocytic hyperplasia where in the thrombocytes are increased with abnormal function. Thrombotic events are seen frequently and represent the main cause of morbidity and mortality in patients with MPNs, mainly polycythemia vera and ET. This study has aimed to research the effects of clonally increased thrombocytes on plasma viscosity (PV) levels among patients with ET and the relationship between PV and thromboembolism history, according to the hypotheses about the effects of PV in thromboembolic events among patients with ET. METHODS: A total of 55 patients were enrolled in the study group, 18 of who had been newly diagnosed with ET according to 2016 World Health Organization criteria and had not previously been treated. 37 of them had already been diagnosed with ET and had been treated. There were 47 healthy volunteers in the control group. 5 cc blood samples were taken from the patients into tubes including an anticoagulant to measure their PV levels. RESULTS: PV of the control group was found to be lower than in the study group and both each patient groups (p < 0.05). No relationship was found between the patient groups in terms of PV (p = 0.404). The mean PV levels of the 16 patients with a history of thromboembolism and the 39 patients with no such history were 2.42±0.17 cP and 2.33±0.20 cP, respectively. The mean PV levels were found to be similar according to their history of thromboembolism in all patient groups and in treated patients (p = 0.572 vs p = 0.991). CONCLUSION: We have found that PV levels were increased in clonally increased thrombocytes in patients with ET when compared with the control group. This is the first study in this field according to our knowledge.

Keywords: Essential thrombocytosis; plasma viscosity; thrombosis

Essential thrombocythemia (ET) is an entity of classic Philadelphia chromosome-negative myeloproliferative neoplasms (MPNs), characterized by thrombocytosis with megakaryocytic hyperplasia where in the thrombocytes are increased with abnormal function [[1]]. Approximately 50%of patients with ET are asymptomatic when they are newly diagnosed. The most frequent symptoms (accounting for approximately 30%) include headache, vision disorders, dizziness, atypical chest pain and microvascular symptoms together with erythromelalgia. Hemorrhagic events are less common than thromboembolic events in patients with ET (2–7%vs 11–25%). Thrombotic events are seen frequently and represent the main cause of morbidity and mortality in patients with Philadelphia chromosome-negative myeloproliferative neoplasms (MPNs), mainly polycythemia vera and ET [[1]]. The pathogenesis of blood clotting activation in these diseases is multifactorial and effects from various abnormalities of platelets, erythrocytes, and leukocytes, as well as dysfunctions of endothelial cells [[2]].

The mutation of the Janus Kinase 2 (JAK2) gene is commonly detected in MPNs. JAK2 is a tyrosine kinase and plays an important role in myelopoiesis. JAK2 V617F gene mutation has been found in almost every patient with polycythemia vera as well as in 50%of patients with ET and primary myelofibrosis. Studies investigating thromboembolism in these diseases have shown that leucocyte counts more than 15×109/L and JAK2 V617F gene mutation have been the risk factors for thrombosis [[3]]. High hemoglobin levels in MPNs, especially polycythemia vera are related with increased viscosity and thromboembolic events [[3], [5]].

There are two dimensions of blood viscosity determination including plasma viscosity and whole blood viscosity. Plasma viscosity (PV) is a major determinant of blood viscosity in circulation and consists of protein-based macromolecules of blood and water. The value of PV can be affected by inflammation and tissue injury because of the changes in plasma proteins; therefore C-reactive protein (CRP), fibrinogen and erythrocyte sedimentation rate (ESR) can increase parallel with PV [[6]]. These two types of viscosity have different effects on different situations. For example, the blood flow of cerebral areas is more affected by PV rather than whole blood viscosity and it is believed that PV is more important in regulating cerebral blood flow than whole blood viscosity [[7]]. In case of hyperviscosity, an increase in thromboembolic events and an increased risk of thrombosis has been shown [[8]]. It has been thought that one reason of the thromboembolic events in patient diagnosed with polycythemia vera is PV [[9]]. In contrast to polycythemia vera, there are no studies on the relationship between PV levels in patients with ET and their risk of thrombo-embolism.

This study has aimed to find out the effects of clonally increased thrombocytes on PV levels among patients with ET and the relationship between PV and thromboembolism history, according to the hypotheses about the effects of PV in thromboembolic events among patients with ET.

The study was conducted at Ankara City Hospital Hematology Department. After the approval of the ethics committee, the patients who came to our hospital's hematology outpatient clinic between July 2020 and August 2020 were included in the study. A total of 55 ET patients were enrolled in the study group, 18 of whom had been newly diagnosed with ET according to 2016 World Health Organization criteria and had not previously been treated. 37 of them had already been diagnosed with ET. These patients had already treated with aspirin and/or hydroxyurea according to their risk classification. Patients using drugs that might affect viscosity, such as lipid-lowering agents; and or who has smoking history, recent surgeries, recently has an infectious condition were excluded. There were 47 healthy volunteers in the control group who had consulted internal medicine outpatient clinics for routine controls. Volunteers who had any systemic diseases, using drugs such as lipid-lowering agents, anti-platelet agents, anti-coagulants, and having smoking history, recent surgery or infectious disease were excluded. The patients diagnosed with ET were evaluated for comorbidities which could affect PV. Patients who had been diagnosed with chronic renal failure, hypertension, coronary artery disease, diabetes mellitus, hyperlipidemia, connective tissue disease, chronic anemia, iron deficiency anemia and infectious diseases were excluded from the study. Furthermore, we excluded anyone who had abnormal ESR, CRP and fibrinogen values from the study and control groups. The patients were divided into two groups, as newly diagnosed and previously treated patients.

5 cc blood samples were taken from the patients, who signed the informed consent form, into tubes including an anticoagulant to measure their PV levels. The samples were taken from newly diagnosed ET patients before treatment and from treated patients during their hospital visits. The samples were kept at –80°C after being centrifuged at 3000 rpm for at least 5 minutes for later examination so as to avoid any mistakes concerning the calibration of the test machine. All the samples were melted and re-centrifuged on the measurement day, then measured at 37°C in a Brookfield DV-II+ Cone Plate Viscometer (Brookfield, Stoughton, MA, USA) machine, which was calibrated with distilled water. Ethics committee approval was received from Ankara City Hospital with the number E1-20-857 and informed consent was taken from all patients.

Distributions of continuous variables were examined by Shapiro Wilk. Age and mean platelet volume (MPV) were given as mean±standard deviation (mean±sd) and other continuous variables were provided as median (min-max). Categorical variables were summarized by frequency (%).

Continuous characteristics were compared between groups by Independent samples t-test, Mann-Whitney U test or Kruskal-Wallis test with respect to the distribution of characteristics and the number of the groups. When necessary, Mann-Whitney U test with Bonferroni correction was applied as post-hoc test after Kruskal-Wallis test. Chi-square tests were performed to compared categorical characteristics between groups. Pearson (r) and Spearman (rS) correlation coefficients were calculated to examine the relationship between PV and age, white blood cell count (WBC), platelet count and hemoglobin level. Binary logistic regression analysis was performed to determine the effect of PV on the history of thromboembolism in ET patients. The enter method was used to construct multivariate model with age, gender, and PV. The odds ratio (OR) and its 95%(CI) were provided as a result. ROC curve analysis was performed to determine the power of platelet count and PV the newly in discriminating the newly diagnosed patients from the control group. Area under curve (AUC) was provided with its standard error (SE). The optimal cut-off points were determined by Youden Index. The Wilson score 95%CIs were calculated for sensitivity and specificity. A p value < 0.05 was considered as statistically significant.

All statistical analyses were performed via IBM SPSS Statistics 21.0 (IBM Corp. Released 2012. IBM SPSS Statistics for Windows, Version 21.0. Armonk, NY: IBM Corp.).

The demographic features and the hematological measurements of the patients are given in Table 1. The mean age of the ET patients was higher than that of the control group (p < 0.001). The mean age of the newly diagnosed and the treated ET patients did not differ (p = 0.055), while the mean age of the treated patients was found to be higher than that of the control group (p < 0.05). The gender distribution across the groups was similar (p = 0.373).

Table 1 Comparisons of demographic features, hematological measurements, and PV across groups

WBC: White blood cell count; MPV: Mean platelet volume; MCH: Mean corpuscular hemoglobin; PV: Plasma viscosity. Bold descriptive statistics are significantly different from other two groups as the result of post-hoc test. 1: comparisons between controls and all patients; 2: comparisons between controls, newly diagnosed and follow-up patients, 3: comparisons between newly diagnosed and follow-up patients. ¥: p < 0.05.

The hemoglobin level in the control group was significantly higher than those in the ET patient group (p = 0.007). The hemoglobin levels of both ET patient groups were similar (p = 0.501, Table 1). The WBC counts of the ET patients were found to be higher than those of the control group (p = 0.028) and the WBC counts of the newly diagnosed patients were significantly higher than those of the treated patients and the control group (p < 0.05, Table 1). The groups were found to be similar in terms of MPV level and ferritin level (p > 0.05, Table 1).

The viscosity level in the control group was significantly lower than those in the ET group and in two patient groups (p < 0.05). In contrast, two patient groups were similar with respect to the viscosity (p = 0.404, Table 1).

The, JAK2, ESR, CRP measurements and thromboembolic-bleeding history of ET patient groups are shown in Table 2. There were 16 (29.1%) ET patients who had a history of thromboembolism and only one of them was a newly diagnosed patient. The history of thromboembolism was found significantly less frequent in the newly diagnosed patients than those in the treated patients (p = 0.018). JAK2 V617F mutation was similar in both ET patient groups (p = 0.055). There was not a significant difference between the groups in terms of their ESR and CRP levels (p > 0.05).

Table 2 Comparisons of JAK2, ESR, CRP measurements and thromboembolic and bleeding history across patient groups

ESR: Erythrocyte sedimentation rate; CRP: C-reactive protein.

The mean PV levels of the 16 patients with a history of thromboembolism and the 39 patients with no history were 2.42±0.17 cP and 2.33±0.20 cP, respectively (Fig. 1). The mean PV levels of treated patients with a history of thromboembolism (n = 15) and without history (n = 22) were 2.35±0.14 cP and 2.35±0.21 cP, respectively (Fig. 1). The mean PV levels were found to be similar according to their history of thromboembolism in all patient groups and in treated patients (p = 0.466 and p = 0.991, respectively).

Graph: Fig. 1 Mean plasma viscosity level with respect to the history of thromboembolism in all patients (right) and in the treated patients (left).

The PV was positively, weakly, and linearly correlated with the age (rS = 0.227, p = 0.022) and the platelet counts (rS = 0.303, p = 0.002), and negatively, weakly, and linearly correlated with hemoglobin (rS = –0.241, p = 0.002) in all individuals. There was no significant relationship between the PV level and age, hemoglobin, and platelet counts in the ET patient groups (p > 0.05 respectively, Table 3).

Table 3 Correlations between PV and age, hemoglobin, and platelet counts

rS: Spearman correlation coefficient. *Pearson correlation coefficient.

The effect of PV on the thromboembolism history in ET group was examined by both a univariate model and a multivariate model considering age and gender of the patients. Neither univariate analysis nor multivariate analysis reveal a significant effect of PV (Table 4).

Table 4 Effect of PV on thromboembolism history

b: Regression coefficient, SE: Standard error, OR: Odds Ratio, CI: Confidence interval. 1 < 60 age and 2Female were reference groups.

The ability of PV levels in terms of separating the newly diagnosed patients from the control group was found to be lower than the platelet count but it was statistically significant (AUC±SE = 0.801±0.059, p < 0.001, Figure 2). The optimal cut-off for PV was determined as≥2.145 cP with 77.8%(95%CI: 54.8%–91.0%) of sensitivity and 74.5%(95%CI: 60.5%–84.8%) specificity.

Graph: Fig. 2 ROC curve of platelet count and plasma viscosity.

Thromboembolic events are the main cause of mortality and morbidity in MPNs [[10]]. Risk factors playing a role in thromboembolism in MPNs are endothelial dysfunction, vascular cell activation, elevated WBC counts, increased PV levels, decreased protein S levels, augmented production of thrombin [[11]].

It was shown that elevated viscosity caused by high hemoglobin levels was the reason for thromboembolic events in polycythemia vera [[3], [5]]. Although some in vitro test studies, which are presumed to be related to plasma coagulation factors decrease, showed deterioration in coagulation tests by increasing hematocrit with the addition of red blood cells, in vivo, increased hematocrit causes hypercoagulability by increasing PV [[12]]. Additional studies have shown that there is a relationship between WBC levels and thrombosis risk among patients with polycythemia vera and ET. Higher WBC levels have been found significantly related with thromboembolic events [[3]]. However there is no study about the relationship between the PV levels of ET patients and the risk of thromboembolism. This study, which we put forward based on the fact that it is not directly researched, shows itself as the first study in this field. We have found that PV levels were increased in clonally increased thrombocytes in patients with ET when compared with the healthy control group. There was no significant difference of PV levels between newly diagnosed patients and treated patients. Moreover, there was no significant relation between PV levels and thrombocyte counts in these two ET patient groups. This could be explained by the fact that there was no cytoreductive therapy indication for most of the patients. 56.3%of the patients in this study had not been receiving cytoreductive therapy.

The effects of thrombocyte count on PV levels had been searched in different studies. Toprak et al. had searched the effects of thrombocyte counts on PV levels. Twenty patients who had thrombocytosis caused by iron deficiency anemia were included in their study. There was no relationship between thrombocyte counts and PV levels. In fact, fibrinogen, hyperlipidemia, and inflammation that could affect PV levels had not been evaluated clinically or laboratory in the study [[14]]. We have also evaluated and compared these parameters to show that there was no difference between the groups in terms of these parameters in order to eliminate the effects of these conditions on PV.

Another study which was conducted among 113 patients with iron deficiency anemia, pancytopenia, polycythemia vera, essential thrombocythemia, idiopathic thrombocytopenic purpura, myelodysplastic syndrome, aplastic anemia, and thalassemia, showed that WBC and platelet counts could affect complete blood viscosity. In contrast they could not show the effect on PV levels [[15]].

Our results were similar to these studies, which were conducted with different ET patient groups. We have found that there was no relationship between thrombocyte count and PV levels in ET patients.

A study which was aimed to investigate the relationship between thromboembolism and PV levels had showed that patients with pulmonary thromboembolism had higher PV levels in than healthy controls [[16]]. Unfortunately, according to the literature so far it cannot be possible to compare the relationship between PV and secondary thromboembolic events caused by different reasons in ET patients who have already had higher PV levels than in the healthy population. In our study, no relationship was found in the groups between PV levels regarding to the history of thromboembolism. This result could be explained as there is not a relation between PV levels and thromboembolic event in ET patients.

JAK2 V617F mutation positivity in our ET patients was 49.1%, thromboembolic event history rate was 29.1%and bleeding history rate in was 5.1%. These results showed a good correlation with literature.

The most important limitation of this study was the number of participants. Another limitation was the difference of age between ET patients and control group, even if we have tried to control the effects of such confusing factors, age remained as a problem.

In conclusion, PV levels were found higher in patients with ET. The relationship between increased plasma viscosity and thromboembolism which plays such an important role in terms of mortality and morbidity among these patients should be searched among larger patient populations.

We would like to thank Prof. Dr. Osman İlhan for giving us permission to use the apheresis unit of Ankara University to measure viscosity.