Abstract
Objectives. The risk of venous thromboembolism (VTE) is increased in patients with antineutrophil cytoplasmic antibodies (ANCA) associated vasculitides (AAV) as compared to healthy subjects. The mechanisms underlying this increased occurrence of VTE are not completely understood. We hypothesize that AAV patients in remission are more procoagulant than healthy controls.
Methods. Patients with AAV in remission and no VTE for the last 6 months were included. Patients with severe renal impairment (serum creatinine > 250 μmol/l) were excluded. Age and sex matched healthy controls were included. The endogenous thrombin potential (ETP) was determined together with hemostatic variables: fibrinogen, D-dimers, factor VIII (FVIII), tissue factor pathway inhibitor (TFPI), protein C, and free protein S.
Results. Thirty-one patients were included. In 27 patients not taking anticoagulants, ETP was measured and found to be elevated: 137.1% as compared to a median of 90.0% for healthy controls (p < 0.01). Fibrinogen and D-dimer levels were not elevated in patients (median 3.5 g/l and 279 μg/l, respectively). FVIII and TFPI levels were also significantly increased in patients as compared to healthy controls (159% vs 137%; 122.5% vs 101%, respectively), whereas protein C and free protein S levels were not elevated (126.5% vs 118.6% and 124.6% vs 118.3%, respectively).
Conclusion. Patients with AAV in remission are more procoagulant than healthy controls, as indicated by an increased ETP. The increased FVIII level measured in these patients suggests persistence of endothelial activation and/or dysfunction. This endothelial dysfunction may cause a continuous low-grade procoagulant state.
Antineutrophil cytoplasmic antibodies (ANCA) associated vasculitides (AAV) belong to the small vessel vasculitides. ANCA in these vasculitides are directed to proteinase-3 (PR3) or myeloperoxidase (MPO), both myeloid enzymes primarily located in the primary granules. Treatment of AAV has improved over past decades, but comorbidities remain significant1. Previous studies demonstrated an increased incidence rate for venous thromboembolism (VTE) in patients with AAV compared to the general population2,3,4,5,6 and not only during active AAV, but also when patients are in remission4.
The mechanisms responsible for the increased incidence of VTE among patients with AAV are not completely understood. Several risk factors, such as genetic susceptibility and/or endothelial cell activation and/or damage, have been suggested7. Further, inflammation induces a procoagulant state8. Since AAV patients also have a higher incidence of VTE during periods when they are in remission, we hypothesize that these patients are hypercoagulable during remission.
MATERIALS AND METHODS
Patients
Consecutive patients who visited the outpatient vasculitis clinic between December 2005 and December 2007 and had a diagnosis of AAV were included in this study9. All patients had either proteinase-3 (PR3) ANCA or myeloperoxidase (MPO) ANCA at the time of diagnosis10. Inclusion criteria were quiescent disease in accordance to European Vasculitis Study Group/European League Against Rheumatism guidelines for definitions of disease activity11. Exclusion criteria were the occurrence of a VTE within 6 months prior to the study and/or severe renal impairment (defined as serum creatinine > 250 μmol/l). We included age and sex matched healthy controls as a control group. Healthy controls were volunteers from the hospital staff without a history of thrombotic events. The local ethics committee approved the study.
Methods
Fasting venous blood samples from patients and controls were drawn at inclusion. By 2-step centrifugation (2000 ×g for 15 min followed by 11,000 ×g for 10 min), platelet poor plasma (PPP) was prepared.
Baseline measurements
At the time of inclusion, additional data were obtained: high sensitivity C-reactive protein (hs-CRP), serum creatinine, leukocyte, and platelet count. Proteinuria was determined in urine samples. In addition, ANCA10 and anticardiolipin antibodies12 were measured. Further, patients were checked for cytomegalovirus (CMV) positivity by microparticle enzyme immunoassay (Abbott).
Endogenous thrombin generation potential
Thrombin generation in PPP was determined using the calibrated automated thrombogram method. Thrombin generation was triggered with 1 pM recombinant tissue factor and 4 μM phospholipid, as described13. In addition, ETP was measured in a separate assay with the presence of 1 nM thrombomodulin14. Three variables were derived from the corrected thrombin generation curves: lag time (minutes) defined as time between coagulation initiation and thrombin formation (1/6 of peak height level); endogenous thrombin potential (ETP, nM*min), represented by the area under the curve; and peak height (nM), the maximal height of the curve. All ETP values and peak values were normalized using normal pooled plasma and expressed in percentage of normal to minimize interassay variation.
Factor VIII (FVIII) activity
FVIII activity was measured using a 1-stage clotting assay according to the manufacturer’s instructions (Dade Behring).
Tissue factor pathway inhibitor (TFPI), protein C, and protein S activities
Full length TFPI and total protein S were measured by ELISA, as described15. Protein C levels were measured in assays with chromogenic substrate performed on a Sysmex CA-7000 automated coagulation analyzer with reagents (Dade Behring). Free protein S was measured using an ELISA following the manufacturer’s instructions (Asserachrom).
Statistics
Continuous variables were checked for normality using the Pearson test and presented as mean ± SD or by median and interquartile range (25th to 75th percentile, IQR), where appropriate, and categorical variables by percentage. Differences in continuous and categorical variables were checked using the independent samples t test or Mann-Whitney U-test and the chi-square or Fisher’s exact test, respectively. Associations between 2 continuous variables were tested using linear regression. p < 0.05 was considered significant.
RESULTS
Thirty-one patients with AAV (mean age 51.4 ± 15.4 yrs, 51.6% male) and 36 age and sex matched healthy controls (mean age 54.8 ± 15.0 years, 50% male) were included in the study. Five patients (16.1%) had had a VTE in the past, of whom 3 were still taking vitamin K antagonists at study inclusion. One patient received vitamin K antagonists for an aortic valve replacement. Eight of the 27 (29.6%) patients did not receive any immunosuppressive medication at the time of inclusion in the study. The other 19 patients received different immunosuppressive medications (Table 1). Two patients included had a relapse 4 and 5 months, respectively, before ETP measurement. All other patients had not had active disease > 1 year before ETP measurement (Appendix 1).
Baseline patient characteristics (n = 31) of antineutrophil cytoplasmic antibodies (ANCA) associated vasculitis patients in remission at the time of endogenous thrombin generation (ETP) potential measurement.
Median D-dimer level was 291 (range 180–463) at inclusion. This was comparable to healthy controls (median 259, range 157–447; p = 0.28). All patients were negative for immunoglobulin (Ig)G and IgM anticardiolipin antibodies. Further, all patients tested negative for anti-β2-GPI antibodies. Eighteen patients (58.1%) were positive for IgG anti-CMV antibodies.
Endogenous thrombin generation potential
ETP was measured in all 27 patients who did not use vitamin K antagonists. The overall ETP in these patients was significantly higher than in healthy controls (median 137.1%, IQR 119.0–160.3 vs 90.0%, IQR 69.1–102.4; p < 0.0001; Table 2; Figure 1). Median normalized peak values in these patients were found to be 193.5% (IQR 150.2–244.9) compared to 81.6% (IQR 63.8–103.5) in healthy controls (p < 0.0001; Table 2). Median lag time of thrombin generation was found to be 7.3 (IQR 6.3–7.7) minutes in patients compared to 7.7 (IQR 6.7–9.3) minutes in healthy controls (p = 0.035).
Endogenous thrombin potential (ETP) characteristics of ANCA associated vasculitis patients in remission (n = 27) and age and sex matched healthy controls (n = 36). ETP could not be measured in 4 patients due to anticoagulant use; median ETP with interquartile ranges (IQR) from patients and healthy controls. ETP was higher in patients (p < 0.0001).
Hematological characteristics of antineutrophil cytoplasmic antibodies (ANCA) associated vasculitis patients (n = 31) and healthy controls (n = 36). Subgroups of PR3-ANCA and MPO-ANCA were analyzed.
No difference in ETP between ANCA subtypes could be detected (p = 0.978 for both ETP and peak values, p = 0.06 for lag times; Table 2). ETP in the presence of thrombomodulin was increased in patients compared to healthy controls (p < 0.0001; Table 2).
CMV status did not influence ETP values (mean ETP 138.7% in patients who were CMV negative compared to a mean ETP of 137.3% in patients who were CMV positive; p = 0.314). There was no difference in ETP between groups of patients who received corticosteroids versus immunosuppressive medication versus none, (median ETP 142.5%, IQR 139.3–174.2 vs 126.7%, IQR 110.6–154.3, vs 142.2%, IQR 130.4–170.7, respectively; p = 0.344; Figure 2).
Endogenous thrombin potential (ETP) characteristics of ANCA associated vasculitis patients (n = 27) divided by the medication used. There was no difference in ETP between different medication groups and/or no immunosuppressive use (p = 0.344). Patients in the ‘none’ group (n = 8) used no immunosuppressive medication. Patients in the ‘prednisone’ group used prednisone alone (n = 3), in combination with azathioprine (n = 7), with mycophenolate mofetil (n = 2) or with methotrexate (n = 3). Patients in the ‘other’ group used mycophenolate mofetil (n = 2) or azathioprine (n = 2).
FVIII activity
Median FVIII activity in patients was 159.0% (IQR 148–210) and was significantly higher than in healthy controls (median 136.9%, IQR 110–158; p < 0.0001; Table 2). FVIII was found to positively correlate with ETP upon linear regression analysis in patients (R2 = 0.15; p = 0.044) and in healthy controls (R2 = 0.33; p < 0.0001).
TFPI, protein C, and protein S activities
TFPI (median 122.5% with IQR 108.8–145.0) was significantly higher in patients compared to healthy controls (median 101.0% with IQR 95.0–139.0; p = 0.025; Table 2).
Median protein C activity was 126.5% (IQR 119.7–145.8), similar to healthy controls (median 118.6%, IQR 112.1–136.4; p = 0.189; Table 2).
Total protein S activity was found to be higher in patients (median 111.3%, IQR 102.7–130.3) than in healthy controls (median 99.8%, IQR 88.2–117.0; p = 0.037). In contrast, free protein S was not higher in patients (median 124.6%, IQR 118.3–140.0) compared to healthy controls (median 118.3%, IQR 108.5–130.8; p = 0.116). Protein S ratios, as calculated by dividing free protein S by total protein S, were 1.1 (IQR 0.9–1.2) in patients and 1.18 (IQR 0.9–1.2) in healthy controls (p = 0.170; Table 2).
Followup of patients
Patients were followed for a median of 5 years after ETP measurement. Two patients had a relapse 4 months after ETP measurement whereas all other patients remained in remission ≥ 1 year afterwards. During a total of 157 patient-years, 4 patients developed a venous thromboembolic event (VTE rate 2.5/100 person-years). One patient developed a VTE during active vasculitis, 1 patient after a total hip replacement, and 2 patients had an idiopathic VTE.
DISCUSSION
The main finding in our study is that patients with AAV in remission have a high thrombin generation potential compared to age and sex matched healthy controls. Thrombin generation is an elegant in vitro method to assess potential thrombogenicity in PPP, and a high ETP in vitro strongly suggests a hypercoagulable state in vivo16. We observed high ETP values both in patients with PR3-AAV and in patients with MPO-AAV. The reduction of ETP in the presence of thrombomodulin was significantly less in patients than in healthy controls. This supports the notion of hypercoagulability in patients where the activated protein C pathway is insufficiently able to suppress thrombin generation17. In addition, higher levels of FVIII were detected in patients when compared to healthy controls, suggesting persistent endothelial dysfunction in AAV patients with quiescent disease.
In line with our findings, Hergesell, et al found elevated von Willebrand factor (vWF) activity in patients with inactive AAV18. Moreover, we found that FVIII activity correlated positively with ETP values. The increased FVIII and vWF activities most probably point towards endothelial cell activation, dysfunction, and/or damage19.
Recently, we found that free protein S, one of the natural coagulation inhibitors, was decreased in SLE patients compared to healthy controls20, possibly contributing to a higher thrombogenicity in SLE. However, our study showed that free protein S levels were not decreased. Other anticoagulant proteins analyzed in our study, such as protein C and TFPI levels, were comparable, and respectively higher in patients compared to healthy controls. This suggests that the anticoagulant effects of TFPI and protein C and protein S in patients with AAV in remission are intact.
During active disease states of AAV, inflammation and other additional factors may contribute to hypercoagulability. Antiplasminogen antibodies have been found in active AAV and may contribute to the higher incidence of VTE in AAV patients during active disease21,22.
We found that 4 patients developed a VTE after ETP measurement. One of these patients developed a VTE during active disease. The VTE rate during followup was 2.5/100 person-years. This rate is much higher than in the general population6.
Most patients in our study (93.5%) had a diagnosis of granulomatous polyangiitis (GPA). We included only 1 patient with MPA and 1 patient with eosinophilic GPA. Therefore, not much can be concluded regarding these latter 2 diseases.
Our study shows that in patients with AAV there is evidence of hypercoagulability events even when patients are in remission, which may explain the protracted risk of VTE during apparently quiescent periods of disease. It may be that different factors are at play during active and inactive disease states. Future studies should further elucidate the complex interactions between the immune system and the coagulation system in AAV, which may facilitate better prophylactic antithrombotic treatment in those at highest risk of VTE.
APPENDIX 1. Additional clinical characteristics of the patients with antineutrophil cytoplasmic antibodies (ANCA) associated vasculitis included (n = 31). Data of diagnosis, as well as disease relapses before and after endogenous thrombin potential (ETP), are reported.
Footnotes
-
Supported by the Dutch Kidney Foundation [KBSO10.010 to MH].
- Accepted for publication August 2, 2013.
REFERENCES
In this issue
Jump to section
- Article
- Abstract
- MATERIALS AND METHODS
- RESULTS
- DISCUSSION
- APPENDIX 1. Additional clinical characteristics of the patients with antineutrophil cytoplasmic antibodies (ANCA) associated vasculitis included (n = 31). Data of diagnosis, as well as disease relapses before and after endogenous thrombin potential (ETP), are reported.
- Footnotes
- REFERENCES
- Figures & Data
- Info & Metrics
- References
- eLetters