SEARCH

SEARCH BY CITATION

Keywords:

  • fibrinogen;
  • platelets;
  • primary immune thrombocytopenia;
  • recombinant factor VIIa;
  • thromboelastometry

Summary

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Disclosure of Conflicts of Interests
  9. References

Haemostatic treatment modalities alternative to platelet transfusion are desirable to control serious acute bleeds in primary immune thrombocytopenia (ITP). This study challenged the hypothesis that recombinant activated factor VII (rFVIIa) combined with fibrinogen concentrate may correct whole blood (WB) clot formation in ITP. Blood from ITP patients (n = 12) was drawn into tubes containing 3·2% citrate and corn trypsin inhibitor 18·3 μg/ml. WB [mean platelet count 22 × 109/l (range 0–42)] was spiked in vitro with buffer, donor platelets (+40 × 109/l), rFVIIa (1 or 4 μg/ml), fibrinogen (1 or 3 mg/ml), or combinations of rFVIIa and fibrinogen. Coagulation profiles were recorded by tissue factor (0·03 pmol/l) activated thromboelastometry. Coagulation in ITP was characterized by a prolonged clotting time (CT, 1490 s (mean)) and a low maximum velocity (MaxVel, 3·4 mm × 100/s) and maximum clot firmness (MCF, 38·2 mm). Fibrinogen showed no haemostatic effect, whereas rFVIIa reduced the CT and increased the MaxVel. The combination of fibrinogen and rFVIIa revealed a significant synergistic effect, improving all parameters (CT 794 s, MaxVel 7·9 mm × 100/s, MCF 50·7 mm) even at very low platelet counts. These data suggest that rFVIIa combined with fibrinogen corrects the coagulopathy of ITP even at very low platelet counts, and may represent an alternative to platelet transfusion.

Primary immune thrombocytopenia (ITP) is an autoimmune disorder characterized by a low platelet count and increased risk of bleeding (Provan et al, 2010). The course of ITP is often benign, but a significant number of patients suffer from serious or even fatal bleeds contributing to an overall reduced life expectancy (Norgaard et al, 2011).

Prophylactic regimens aim at achieving a stable platelet count to secure haemostasis by applying the least harmful treatment possible. Occasionally, patients may require an immediate improvement of the haemostatic capacity due to serious active bleeding or urgent need for invasive procedures (Provan et al, 2010). However, prompt haemostatic management still relies on extensive platelet transfusions as the only recognized treatment option.

Transfusion of donor platelets has been found useful in emergency treatment of ITP by retrospective evaluations (Carr et al, 1986; Spahr & Rodgers, 2008). Furthermore, massive platelet transfusion was found invariably effective in a group of 10 patients (Salama et al, 2008). However, treatment with platelets may be ineffective, delayed, or associated with serious adverse events, such as platelet antibodies, allergic reactions, sepsis, transfusion-related acute lung injury, and possibly increased mortality (Spiess et al, 2004; Keller-Stanislawski et al, 2009). Hence, haemostatic treatment modalities, circumventing the need for platelet transfusion, would be highly valuable in ITP. In our consideration, potential candidates may include recombinant activated factor VII (rFVIIa; NovoSeven®, Novo Nordisk, Bagsværd, Denmark; Logan & Goodnough, 2010) and fibrinogen concentrate (Riastap®/Haemocomplettan®; CSL Behring, Marburg, Germany; Fenger-Eriksen et al, 2009a).

A summary of numerous case reports suggested the usefulness of rFVIIa in controlling haemorrhage in patients with severe ITP (Salama et al, 2009). In addition, a haemostatic effect of rFVIIa in thrombocytopenia is supported by clinical (Kristensen et al, 1996), laboratory (Kjalke et al, 2001; Lisman et al, 2005; Larsen et al, 2010) and animal studies (Lauritzen et al, 2009). However, a randomized trial of bleeding patients with thrombocytopenia following haematopoietic stem cell transplantation showed no overall significant effect of rFVIIa on bleeding score (Pihusch et al, 2005), and no randomized trials have been performed in immune-mediated thrombocytopenia. Hence, the efficacy of rFVIIa in thrombocytopenia is still unresolved (Logan & Goodnough, 2010), and rFVIIa cannot be recommended as a standard treatment for bleeding episodes.

Fibrinogen acts as the bridging molecule between activated platelets and is the crucial substrate in the construction of a dense fibrin network (Fenger-Eriksen et al, 2009a). Importantly, a pioneering experimental study (Velik-Salchner et al, 2007) suggested a haemostatic effect of supraphysiological levels of fibrinogen in thrombocytopenia, and these observations have been supported by laboratory studies (Lang et al, 2009).

Combining rFVIIa and fibrinogen has been suggested to be beneficial in ex vivo studies on coagulopathies associated with cardiopulmonary bypass surgery (Tanaka et al, 2008; Sorensen et al, 2009) as well as dilutional coagulopathy (Ganter et al, 2008; Fenger-Eriksen et al, 2009b). Hence, it may be speculated, that enhanced thrombin generation stimulated by rFVIIa (Kjalke et al, 2001) together with improved fibrin polymerization following fibrinogen substitution (Mosesson, 2005; Lang et al, 2009) would be beneficial in isolated thrombocytopenia.

The incidence and severity of bleeding manifestations in ITP are largely associated with the platelet count (Lacey & Penner, 1977). The haemostatic efficacy of rFVIIa in thrombocytopenia also appears to be affected by the platelet count (Kristensen et al, 1996; Monroe et al, 1997) and function (Larsen et al, 2009). Nevertheless, systematic reports on the correlation between the haemostatic response to potential interventions and the platelet count are still lacking.

Hence, this study aimed to (i) characterize the coagulopathy of ITP; (ii) investigate the in vitro haemostatic response to rFVIIa, fibrinogen, and the combination of rFVIIa and fibrinogen; and (iii) investigate the correlation between the haemostatic response and the baseline platelet counts.

Methods

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Disclosure of Conflicts of Interests
  9. References

Study subjects

The study was approved by appropriate Research Ethics Committees in Denmark and the UK. Following written informed consent, blood samples were drawn from 12 patients diagnosed with ITP, five males and seven females with a mean age of 53 years (range 30–85 years). The patients had a mean platelet count of 22 × 109/l (range 0–42), elevated immature platelet fractions (mean: 18·4%, range 4·7–31·8), normal levels of fibrinogen (3·1 g/l, range 2·4–4·1), and none displayed von Willebrand Factor deficiency. Follow up was performed in three patients commencing conventional treatment for low platelet counts. Blood was also drawn from 15 healthy adults, seven males and eight females, with a mean age of 28 years (range 25–34), for healthy reference and for generation of supportive thrombocytopenia model data (Larsen et al, 2007). All participants had not received drugs known to affect platelet function or coagulation, and all displayed prothrombin times (PTs) and activated partial thromboplastin times (APTTs) within the normal range.

The clinical phenotype of the ITP patients was assessed according to an ITP bleeding score (Page et al, 2007) by a single physician. Signs of bleeding were observed with the following frequency at first blood sampling: Skin 75%, oral 17%, whereas the frequency of bleeding symptoms within the last week prior to first blood sampling were reported as: skin 83%, oral 33%, epistaxis 25%, gynaecological 17%, gastrointestinal 8%, and pulmonary 8%, with 58% of the patients reporting bleeding symptoms from multiple sites.

Reagents and buffers

HEPES buffer (VWR, Herlev, Denmark) was used as buffer control. Corn trypsin inhibitor (CTI) was purchased from Haematologic Technologies, Essex Junction, VT, USA. Recombinant human tissue factor (TF; Innovin®; Dade Behring, Marburg, Germany) was diluted in HEPES buffer containing 200 mmol/l CaCl2, and used as activator of whole blood (WB) clot formation. The Innovin® TF activity was determined using the Actichrome® TF chromogenic activity assay (American Diagnostica Inc., Greenwich, CT, USA). For evaluation of thrombin generation, another TF source (PRP-reagent, final concentration 1 pmol/l), FluCa-kit and Thrombin Calibrator were obtained from Thrombinoscope BV (Maastricht, the Netherlands). Recombinant FVIIa (NovoSeven®) and fibrinogen concentrate (Riastap®/Haemocomplettan®) were obtained from respective manufacturers. Horm® collagen was from Nycomed, Linz, Austria.

Blood sampling

Blood intended for WB thromboelastometry and thrombin generation was drawn directly into 0·106 mol/l trisodium citrate (S-Monovette® tubes; Sarstedt, Nümbrecht, Germany) containing CTI at a final concentration of 18·3 μg/ml in order to inhibit artificial contact pathway activation (Luddington & Baglin, 2004). No CTI was added to the blood used for platelet aggregation and standard analyses.

Laboratory procedure for generating thrombocytopenia in healthy whole blood

Severe thrombocytopenia was produced in WB from the healthy volunteers by applying a validated laboratory model (Larsen et al, 2007).

Whole blood coagulation analyses

Dynamic WB clot formation profiles were recorded by ROTEM® Thromboelastometry (Tem International, Munich, Germany) using a previously described methodology (Sorensen et al, 2003). In brief, the citrate-CTI WB rested at ambient temperature for 30 min and reaction mixtures were prepared mixing 300 or 280 μl of WB with 20 μl buffer/rFVIIa/fibrinogen or 40 μl of platelet-rich plasma (PRP), respectively. The WB of ITP patients was spiked with buffer, PRP from a single, well-characterized donor (increasing the platelet count by 40 × 109/l, representing the potential effect of platelet concentrates), rFVIIa (1 or 4 μg/ml), fibrinogen (1 or 3 mg/ml) as well as combinations of fibrinogen and rFVIIa. The doses roughly corresponded to the administration of a standard dose (rFVIIa 90 μg/kg, fibrinogen 70 mg/kg) in line with clinical practice (Fenger-Eriksen et al, 2009a; Salama et al, 2009), as well as a high dose. All analyses were initiated by addition of 20 μl of TF (final concentration 0·03 pmol/l) and CaCl2 based on titration experiments in thrombocytopenia model blood. Parameters evaluating clot initiation (clotting time, CT) and final clot strength (maximum clot firmness, MCF) were recorded in addition to dynamic parameters of the propagation phase, such as the maximum velocity (MaxVel) and time to maximum velocity (t,MaxVel) of clot formation (Sorensen et al, 2003). All results are given as means of duplicate experiments, and a total run-time of 90 min was applied.

Thrombin generation by calibrated automated thrombography

The assay was performed according to the manufacturer (Thrombinoscope BV), essentially as devised by Hemker et al (2003). However, to reflect the thrombocytopenic condition, platelet counts were adjusted to 10 and 30 × 109/l, when sufficient platelets were present in patient PRP. Hence, the analysis was performed on n = 9 and n = 6, respectively. The intervention procedures followed an identical pattern as described for WB clot formation experiments.

Whole blood platelet aggregation

Platelet aggregation in ITP WB was investigated by Multiplate® impedance aggregometry (Verum Diagnostica, Munich, Germany). Citrated WB (300 μl) and isotonic saline (270 μl) was added to the reaction cup and incubated at 37°C for 3 min. The aggregation was initiated by addition of 30 μl of collagen (final concentration 8 μg/ml), which was the lowest final concentration providing a complete aggregation amplitude as determined by titration experiments (Larsen et al, 2009).

Statistical considerations

Statistical analyses were performed using stata 11·2 (StataCorp LP, College Station, TX, USA). Data were well described by a Gaussian distribution and are reported as means ± standard deviations (SD). Differences were evaluated by a paired Student's t-test or a two-sample t-test allowing unequal variances as appropriate. Simple linear regression analyses were performed using the baseline thromboelastometric data or the differences observed after addition of the various interventions (intervention – baseline) as the dependent variable (y) and the platelet count as the independent variable (x). The slope was used to evaluate dependency of the haemostatic response on the platelet count, whereas the intercept was used to evaluate the haemostatic response at very low platelet counts. P-values less than 0·05 were assigned significance.

Results

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Disclosure of Conflicts of Interests
  9. References

Whole blood thromboelastometric profiles in baseline ITP and after in vitro interventions

Compared to healthy controls the WB coagulation profiles of the ITP patients were characterized by a prolonged CT (< 0·001) and t,MaxVel (< 0·001) as well as a markedly reduced MaxVel (< 0·001) and MCF (= 0·01; Table 1, Fig 1). Simple linear regression showed a positive correlation between the platelet count and the baseline MaxVel (< 0·001, R2 = 0·76) as well as MCF (P < 0·001, R2 = 0·78).

Table 1. Parameters of whole blood clot formation in primary immune thrombocytopenia
 Primary immune thrombocytopenia (n = 12) Healthy controls (n = 15)
BufferPRP (+40 × 109/l)rFVIIa (1 μg/ml)rFVIIa (4 μg/ml)Fibrinogen (1 mg/ml)Fibrinogen (3 mg/ml)rFVIIa (1 μg/ml) Fibrinogen (1 mg/ml)rFVIIa (4 μg/ml) Fibrinogen (1 mg/ml)rFVIIa (1 μg/ml) Fibrinogen (3 mg/ml)rFVIIa (4 μg/ml) Fibrinogen (3 mg/ml) Buffer
  1. PRP, platelet-rich plasma; rFVIIa, recombinant activated factor VII; CT, clotting time; MaxVel, maximum velocity; t,MaxVel, time to maximum velocity; MCF, maximum clot firmness.

  2. Values are reported as mean and standard deviation (SD).

  3. Primary immune thrombocytopenia versus healthy control samples: †< 0·05, †††< 0·001.

  4. Interventions versus buffer in primary immune thrombocytopenia: *< 0·05, **< 0·01, ***< 0·001.

Clot initiation
 CT (s)1490†††1164**905***744***15701764*957***753***1057***794*** 941
SD293262185179387448221160248186 207
Clot propagation
 MaxVel (mm × 100/s)3·43†††6·96***5·04***6·28***3·192·455·24***6·46***5·64**7·90*** 9·71
SD1·602·492·713·111·581·372·502·792·902·83 2·75
t,MaxVel (s)2049†††1611*1353***1064***204824141329***1090***1459***1177*** 1294
SD390329306190322576245169317227 237
Clot termination
 MCF (mm)38·249·6***38·638·838·539·339·739·548·9***50·7*** 49·4
SD12·57·612·813·214·216·011·012·012·911·2 7·0
image

Figure 1. Representative velocity profiles of whole blood thromboelastometry. (A) Healthy control (solid line) and baseline primary immune thrombocytopenia (ITP; dotted line). (B) Healthy control and ITP before and after haemostatic interventions. The combination of rFVIIa and fibrinogen displayed a synergistic effect reversing the coagulopathy of ITP.

Download figure to PowerPoint

The addition of PRP shortened the CT and t,MaxVel, and increased the MaxVel and MCF (Table 1, Fig 1). Addition of rFVIIa led to a dose-dependent shortening in the CT and t,MaxVel as well as a dose-dependent increase in the MaxVel (all < 0·001; Table 1, Fig 1). The response to rFVIIa (rFVIIa-baseline) observed in the MaxVel was highly correlated with the platelet count (1 and 4 μg/ml: R2 = 0·81 and R2 = 0·86, respectively, < 0·001), and in addition, the estimated response at very low platelet counts (intercept at platelet count 0 × 109/l) was not significant (Fig 2B). However, the estimated intercepts supported a significant improvement in the CT [1 μg/ml: −643s (95% confidence interval CI −881 to −405); 4 μg/ml: −811 s (95% CI −1036 to −586)] and similarly in the t,MaxVel at very low platelet counts (Fig 2A).

image

Figure 2. Haemostatic response (intervention-baseline) versus the platelet count in patients with primary immune thrombocytopenia. (A–C) Development in clotting time, maximum velocity, and maximum clot firmness after addition of rFVIIa 1 μg/ml (open circles, solid regression line (OC)), rFVIIa 4 μg/ml (open squares, short dashed regression line (OS)), and fibrinogen 3 mg/ml (open triangles, long dashed regression line). (D–F) Development following addition of PRP (+40 × 109/l) (OC) and the combination of fibrinogen 3 mg/ml and rFVIIa 4 μg/ml (OS).

Download figure to PowerPoint

Addition of fibrinogen resulted in no positive changes in the coagulation profiles. In contrast, fibrinogen showed a trend to prolong the CT (= 0·046) and t,MaxVel (= 0·081) as well as reduce the MaxVel (= 0·058) (Table 1, Fig 1, Fig 2–C).

The combination of rFVIIa and fibrinogen resulted in a synergistic effect on the MaxVel (combination versus baseline or monotherapy, all ≤ 0·006) and the MCF, where the combinations using fibrinogen 3 mg/ml showed a marked increase (< 0·001, Table 1, Fig 1). In addition, combining rFVIIa 4 μg/ml and fibrinogen 3 mg/ml provided a significant intercept in the MaxVel [3·29 mm × 100/s (95% CI 1·39–5·19)], and the response was not significantly correlated to the platelet count (R2 = 0·21, = 0·13). No significant correlations to the platelet count were observed in the MCF (all P > 0·23), and the combinations containing fibrinogen 3 mg/ml provided significant positive intercepts [with rFVIIa 4 μg/ml: 13·9 mm (95%CI 10·3–17·6); (Fig 2–F].

When comparing the combination of rFVIIa 4 μg/ml and fibrinogen 3 mg/ml to the normal control samples no significant differences were observed in any parameter (CT = 0·06; MaxVel = 0·11; t,MaxVel = 0·20; MCF = 0·74; Table 1). In contrast, raising the platelet count by 40 × 109/l only normalized the MCF (= 0·94).

Thromboelastometric profiles in whole blood thrombocytopenia model before and after in vitro interventions

Following induction of thrombocytopenia in WB from healthy volunteers a platelet count of 25 × 109/l (range 6–52) was obtained. As in the ITP patients, the WB thrombocytopenia model was characterized by a reduced MaxVel (5·3 mm × 100/s, < 0·001) and MCF (33·4 mm, < 0·001), whereas no significant changes were observed in the CT (876 s) or t,MaxVel (1195 s) when compared to the healthy control samples.

Addition experiments with rFVIIa as well as the combination of rFVIIa and fibrinogen supported the findings in ITP patients. In contrast, fibrinogen showed a significant positive effect on the MaxVel (7·2 mm × 100/s) and MCF (39·0 mm; < 0·001).

Thrombin generation in ITP ‘platelet-rich’ plasma

The baseline thrombin peak was reduced in ITP plasma with platelet counts adjusted to 10 × 109/l (30·7 ± 11·4 nmol/l) and 30 × 109/l (49·2 ± 5·9 nmol/l) compared to controls performed in each experiment at 150 × 109/l (107·9 ± 11·8 nmol/l). As illustrated in Fig 3, the response patterns to rFVIIa and fibrinogen were similar when platelet counts were adjusted to 10 or 30 × 109/l, and data from platelet counts adjusted to 10 × 109/l are presented in the following. The addition of fibrinogen 3 mg/ml resulted in a decrease in thrombin peak (< 0·001; Fig 3), as well as a prolongation of the lag-time (P < 0·001). In contrast, rFVIIa increased the thrombin peak and reduced the lag-time considerably (all < 0·001), thereby off-setting the negative effect of supraphysiological fibrinogen on thrombin generation when used in combination (Fig 3). Parallel investigations in healthy control samples (n = 3) at platelet counts adjusted to 30 × 109/l showed a similar pattern of response to the interventions, but a considerably higher baseline level of thrombin peak (65·2 vs. 49·2 nmol/l) as well as a shorter baseline lag-time (8·3 vs. 9·0 min).

image

Figure 3. Development in thrombin peak following in vitro addition of fibrinogen and rFVIIa. Platelet count adjusted 10 × 109/l (n = 9, solid lines), and platelet count adjusted to 30 × 109/l (n = 6, dashed lines) in plasma from patients with primary immune thrombocytopenia. The plots show no addition of rFVIIa (open circles), and addition of rFVIIa 1 μg/ml (open squares) and 4 μg/ml (open triangles). The symbols denote mean and standard error of mean (error bars).

Download figure to PowerPoint

Whole blood platelet aggregation in ITP

The ITP patients displayed markedly increased platelet aggregation compared to previously published data on patients with malignancy-related thrombocytopenia with similar platelet counts (64·7 ± 52·0 aggregation units (AU) vs. 9·1 ± 4·9 AU, = 0·004; Fig 4; Larsen et al, 2009). However, the aggregation of the WB thrombocytopenia model was not significantly different from the ITP patients (38·9 ± 24·3 AU, = 0·14; Fig 4).

image

Figure 4. Whole blood platelet aggregation in response to 8 μg/ml collagen. The primary immune thrombocytopenia (ITP) patient group (n = 12) compared with previously published data of aggregation in the whole blood thrombocytopenia model (n = 12) and children with thrombocytopenia due to chemotherapy (= 8) (Larsen et al, 2009). The mean platelet counts (citrate) of the groups were 20, 19 and 26 × 109/l, respectively. The bars display mean and standard error of the mean (error bars). ** < 0·01 by unpaired t-test. NS indicates not significant.

Download figure to PowerPoint

Haemostatic improvement in ITP patients during conventional therapy

Three male ITP patients were followed during regular treatment for low platelet counts. Patient A was studied during the acute onset of ITP and initially treated with dexamethasone 40 mg daily and from day 5 concomitantly with intravenous immunoglobulin (IVIg) 1 g/kg per d until finally responding to treatment after 7 d with cessation of bleeding manifestations: skin bruising, petechia, oral blood blisters, epistaxis, macroscopic haematuria, and fresh blood in stools. Platelet counts were: 0, 0, 1, and 68 × 109/l at baseline and 2, 5 and 7 d, respectively. Patient B was started on romiplostim 1 μg/kg for the first week increasing to 2 and 3 μg/kg in the second and third week, respectively. This regimen resulted in steadily increased platelet counts, from 4 × 109/l to 14 × 109/l and 37 × 109/l after 9 and 22 d of treatment. In parallel, outbreaks of skin bruising and oral blood blisters ceased. Patient C received IVIg (1 g/kg per d for 2 d) and responded with a minor increment in the platelet count, from 7 × 109/l at baseline to 21 × 109/l after 5 d, and no development of new skin manifestations and nasal bleeds. The increment in platelet counts and loss of bleeding manifestations during treatment were paralleled by improvement in the CT, MaxVel and MCF in all of the patients (Fig 5). No trends were observed in the development of fibrinogen levels, von Willebrand factor activity, APTT, or PT.

image

Figure 5. Development in whole blood clot formation in three primary immune thrombocytopenia patients during conventional treatment. The parameters: clotting time (solid lines), maximum velocity (dotted lines) and maximum clot firmness (dashed lines) are displayed for Patient A (open circles), Patient B (open squares), and Patient C (open triangles). The improvements in thromboelastometric parameters concurred with reduced bleeding manifestations as well as increments in platelet count.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Disclosure of Conflicts of Interests
  9. References

The main findings of the present study are that the combination of rFVIIa and fibrinogen shows a synergistic effect on WB clot formation correcting the coagulopathy of ITP and, of note, provides a significant haemostatic response even at very low platelet counts. Importantly, fibrinogen alone did not provide a haemostatic effect, and the response to rFVIIa was highly dependent on the platelet count.

Baseline ITP coagulation profiles were characterized by a prolonged clot initiation, a reduced clot propagation and clot strength, as well as reduced thrombin generation. These observations are consistent with previous findings (Kjalke et al, 2001; Lang et al, 2009; Larsen et al, 2009).

The ITP patients displayed normal levels of fibrinogen, and supplemental addition of fibrinogen appeared to cause a dose-dependent inhibitory effect on thrombin generation and no positive effect on clot formation. In contrast, a positive effect of fibrinogen was observed on clot formation in the WB thrombocytopenia model also displaying increased thrombin generation when compared to the ITP samples, possibly due to substances in the added PPP affecting thrombin generation (Cattaneo et al, 2007). The inhibitory effect of fibrinogen/fibrin on thrombin generation is well established but not widely appreciated (Mosesson, 2005). The present observations emphasize the importance of sufficient thrombin generation in order to utilize the extra added fibrinogen, suggesting that thrombin generation in ITP patients with low platelet counts is insufficient to benefit from additional fibrinogen. Velik-Salchner et al (2007) reported a significant decrease in bleeding following fibrinogen in thrombocytopenic pigs. This is compatible with our findings, as pigs are hypercoagulable with enhanced thrombin generation compared to humans (Velik-Salchner et al, 2006). Previous in vitro experiments on human blood showed improved WB clot formation following fibrinogen addition in thrombocytopenia (Lang et al, 2009; Misgav et al, 2011). However, these studies did not use CTI and the positive effect of fibrinogen may be an artefact due to fibrin polymerization driven by contact-pathway-determined generation of thrombin as well as non-physiological high levels of TF applied in standard thromboelastometry assays (Mann et al, 2007; Larsen et al, 2011).

Recombinant FVIIa provided a dose-dependent effect on clot initiation and a reduction in lag-time of thrombin generation, which corresponds well to previous findings where rFVIIa particularly enhanced the initial thrombin generation and platelet activation in thrombocytopenia-like conditions (Kjalke et al, 2001). In contrast to previous findings in malignancy-related thrombocytopenia (Larsen et al, 2009), rFVIIa also improved the propagation phase and increased the thrombin generation. This disparity may partly be explained by the augmented platelet function in ITP compared to malignancy-related thrombocytopenia demonstrated by our data in line with others (Misgav et al, 2011). In addition, the present study utilized CTI to abolish artificial contact-pathway activation (Rand et al, 1996), which has previously been shown to mask the haemostatic response to rFVIIa (Larsen et al, 2010).

The present study did not attempt to investigate the in vivo clinical correlation of the ex vivo effect of the interventions. However, it seems reasonable to assume that normalization of all aspects of clot formation would be beneficial when aiming to achieve haemostasis in ITP. Recombinant FVIIa did not normalize clot formation, however, some positive changes were observed following addition. These modest findings may reflect the clinical observations suggesting that some bleeding thrombocytopenic patients might benefit from infusion of rFVIIa (Kristensen et al, 1996; Salama et al, 2009), whereas others clearly do not (Pihusch et al, 2005). Considering the present evidence, rFVIIa cannot be recommended as a standard treatment for bleeding episodes in thrombocytopenia.

Importantly, combined administration of rFVIIa and fibrinogen provided a synergistic effect on WB clot formation resulting in profiles similar to those obtained from healthy controls, and a significant improvement in all haemostatic characteristics of clot formation, even at very low platelet counts. The advantages of combining rFVIIa and fibrinogen observed in our study corroborates well with previous in vitro findings in other patient groups (Ganter et al, 2008; Tanaka et al, 2008; Fenger-Eriksen et al, 2009b; Sorensen et al, 2009).

The applied low-TF thromboelastometric assay was able to illustrate the improvements in the haemostatic capacity that occurred in parallel with the clinical improvements attributed to the gain in platelet counts observed in three ITP patients undergoing conventional treatment. In our opinion these observations support the validity of the method. However, several study limitations merit attention. Most importantly, the present study is an in vitro investigation and any extrapolation of the findings into the clinical setting should be done with great caution. Therefore, we emphasize that clinical safety as well as dose ranging studies are needed before any clinical application can be recommended. Furthermore, shortcomings of thromboelastometry mimicking in vivo clot formation include lack of flow as well as interactions with blood vessel surface and subendothelial tissue. Potential powerful sources of thrombin, such as prothrombin complex concentrate and activated prothrombin complex concentrate, may be alternatives to rFVIIa, but were not investigated in this study.

In conclusion, the coagulopathy of ITP was characterized by delayed and compromised WB clot formation as well as decreased thrombin generation. Recombinant FVIIa combined with fibrinogen corrected clot formation in ITP even at very low platelet counts, and may be suggested as an alternative to platelet transfusion. The proposed pharmacological mechanism includes local stimulation of thrombin generation by rFVIIa, possibly facilitated by a favourable platelet function in ITP patients, together with enhancing three-dimensional clot formation by supplementation with fibrinogen. Importantly, fibrinogen alone seems unlikely to be useful and may even provide an anti-thrombin effect. Evidently the mechanism needs further evaluation and, ultimately, clinical trials would be preferable to further investigate if this new treatment modality holds the potential to serve as an effective acute treatment option in ITP.

Acknowledgements

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Disclosure of Conflicts of Interests
  9. References

The authors would like to thank research technician Kirsten Christiansen for highly qualified assistance. This work was supported by unrestricted grants from Aarhus University, Helga and Peter Korning's Foundation, and The A.P. Møller Foundation for the Advancement of Medical Sciences.

Author contributions

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Disclosure of Conflicts of Interests
  9. References

OHL, JS, JI, and BS designed the study. OHL performed research and data analyses. All authors interpreted the data. OHL wrote first draft of the manuscript. All authors revised the manuscript and approved the final version.

Disclosure of Conflicts of Interests

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Disclosure of Conflicts of Interests
  9. References

BS has participated in advisory committees and/or received speakers' bureau from Novo Nordisk, Baxter, CSL Behring, Bayer, SOBI, and TEM International. The Haemostasis Research Unit has received unrestricted research support from Novo Nordisk, Baxter, CSL Behring, Bayer, SOBI, Grifols, LFB, and Octapharma.

References

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Disclosure of Conflicts of Interests
  9. References