Therapeutic monitoring of von Willebrand disease: interest and limits of a platelet function analyser at high shear rates

Authors


Professor Dominique Meyer, INSERM U.143, 84 rue du Général Leclerc, 94276 Le Kremlin-Bicêtre Cedex, France.

Abstract

We have evaluated the position of the Platelet Function Analyzer PFA-100TM in the management of 41 patients with von Willebrand disease (VWD) receiving either desmopressin (23 patients with type 1, five with type 2M, three with type 2A and three with type 2B) or von Willebrand factor (VWF) concentrates (four patients with type 3, two with type 2M ‘type B’, two with type 2A and one type 1 ‘platelet low’). In all patients the following were studied before and 30 min after infusion of desmopressin and/or VWF concentrates: VWF ristocetin cofactor activity (VWFRCo), bleeding time (BT) and closure time with the PFA-100 using ADP (CT-ADP) as well as epinephrine (CT-Epi) cartridges. After the infusion of desmopressin, the CT was modified in the same way as the VWFRCo levels, being always normalized in patients with type 1 and not constantly corrected in those with type 2. Thus, our results indicated that the measurement of the CT enabled a quick and accurate evaluation of the response to desmopressin which, in fact, measured the releasable VWF cellular compartment containing the highly multimerized forms of VWF. For patients with type 2 or 3 VWD who were non-responsive to desmopressin, VWF concentrates corrected the VWFRCo defect but not the CT as none of these patients had a normal platelet VWF content and the VWF concentrates did not contain the ultralarge VWF multimers. In conclusion, the very high shear conditions in the PFA-100 make it very sensitive to the contribution of platelet VWF and to the ultralarge VWF multimers, indicating that the evaluation of the CT is a very simple and rapid tool to discriminate between good and non-responders to desmopressin.

von Willebrand disease (VWD) (Nichols & Ginsburg, 1997) is the most common inherited bleeding disorder, with a prevalence estimated as high as 1–2% of the population (Rodeghiero et al, 1987). VWD arises from quantitative or qualitative abnormalities of von Willebrand factor (VWF). The current classification (Sadler, 1994) distinguishes partial (type 1) or complete (type 3) quantitative deficiencies from qualitative defects (type 2). Although inheritance is usually autosomal dominant, some variants may be recessively inherited. Type 1, accounting for 70–75% of all cases of VWD, is defined by a simple quantitative deficiency of plasma VWF with normal VWF multimeric structure and function, and variable (low or more often normal) VWF levels in platelets. Type 3 is characterized by a virtually complete deficiency of VWF in plasma and in platelets. The type 2 variants include four main subtypes. Type 2A refers to patients with decreased affinity for platelets associated with the absence of large VWF multimers in plasma (and in some cases in platelets). Type 2M is easily confused with type 1 VWD as there is no loss of high MW multimers but the binding of VWF to platelets is decreased. Type 2B variant is characterized by an increased affinity of VWF for platelet glycoprotein Ib (GPIb) and a variable multimeric distribution of VWF in plasma. In type 2N VWD, patients show a decreased affinity of VWF for FVIII.

In fact, patients with VWD represent an heterogenous group whose clinical symptoms are not specific and vary in severity. The laboratory diagnosis of VWD may be difficult in mild forms, where the screening tests, e.g. bleeding time (BT) and activated partial thromboplastin time (aPTT), may be in the normal range. Indeed, the BT is neither reproducible nor sensitive. The aPTT is prolonged if the FVIII level is sufficiently decreased but is normal in many cases of VWD. It has thus been necessary for the screening of VWD to make use of the ristocetin cofactor assay (VWFRCo), a time-consuming test requiring the contribution of an experienced technologist.

Recently, we clearly showed that the platelet function analyser PFA-100TM (Dade Behring Inc., Deerfield, Ill.) provides a very simple and sensitive routine test for the screening of patients with all types of VWD (except type 2N where the interaction of VWF with platelets is normal) (Fressinaud et al, 1998). We report here on the performance of the PFA-100 in the therapeutic monitoring of patients with VWD.

The optimal therapeutic choice for therapy in patients with VWD is critically dependent on their accurate diagnosis and subclassification. There are two main therapeutic ways: desmopressin infusion and replacement therapy with VWF concentrates. In most patients with VWD, baseline FVIII levels are > 30 IU/dl and the principal aim is to correct the primary haemostatic defect. Full correction of the Ivy BT appears not to be necessary to achieve haemostasis except in bleeding episodes involving mucosal tissues or the central nervous system. Therefore the correction of the VWFRCo levels to > 50 IU/dl seems to be a rational therapeutic goal.

Desmopressin stimulates the release of VWF from vascular endothelial cells (Mannucci et al, 1977; Mannucci, 1997). It is therefore the treatment of choice for most patients with type 1 VWD because it normalizes the BT and the plasmatic VWF levels. In fact, patient responses positively correlate with the platelet VWF levels. The effect of desmopressin is variable in type 2A, generally effective in the group of patients having a normal intracellular synthesis and processing of VWF. In some patients with type 2M desmopressin may be effective. In type 2B VWD the use of desmopressin, which releases the HMW multimers with increased affinity for GPIb and may exacerbate the thrombocytopenia, is considered to be contraindicated although this is controversial (Castaman & Rodeghiero, 1996). In type 3, where no VWF is synthesized, desmopressin is ineffective. In all VWD patients, VWF concentrates, which never contain the full range of plasma VWF multimers, increase the plasma VWF levels in proportion to the quantity infused. However, in patients with type 3 the improvement in the BT is usually only partial and transient (Rodeghiero et al, 1992).

In order to estimate the performance of the PFA-100 in the therapeutic monitoring of VWD we have measured the Ivy BT, the VWFRCo levels and the closure time of the PFA-100 before and after the intravenous infusion of desmopressin or VWF concentrates in patients with different types of VWD.

MATERIALS AND METHODS

Patients

41 patients (13 males and 28 females, aged between 2.5 and 63 years) with congenital VWD were included in the study following appropriate consent. VWD was diagnosed by a personal and/or familial bleeding history and results of laboratory tests such as BT, evaluation of the closure time (CT) with the PFA-100, platelet count, plasma VWF antigen (VWFAg) and VWFRCo assays, FVIII activity, ristocetin-induced platelet aggregation (RIPA) and multimeric analysis of VWF. Patients were classified according to the revised classification of VWD (Sadler, 1994). In some patients we studied platelet VWF (VWFAg and VWFRCo assays and multimeric analysis). In all patients with type 2 we performed plasma VWF binding assay to platelet GPIb (induced in the presence of ristocetin and botrocetin), binding assays to collagen and to FVIII, and the DNA analysis allowed to identify responsible mutations. Therefore we studied 23 patients with type 1, six with type 2M, four with type 2A, three with type 2B, four with type 3 and one patient with type 1 who had the rare ‘platelet-low’ subtype (Mannucci et al, 1985). Two patients with type 2M were sisters and exhibited the particular variant described as ‘type B’ (Rabinowitz et al, 1992).

Desmopressin

Desmopressin (1-deamino-8-d-arginine vasopressin, abbreviated DDAVP) (Minirin®, Ferring S.A., France) at a dose of 0.3 μg/kg body weight in 50 ml saline was administered intravenously over a 30 min period. This was done routinely in order to evaluate the response of VWD patients to DDAVP and to select good and poor responders. Patients were fully aware of the purpose of this procedure. All patients with type 3 known to be unresponsive to DDAVP and three patients with respectively type 2A, type 2M and type 1 ‘platelet low’ in whom desmopressin had already been tested and shown no efficacy, did not receive the drug. Blood was collected before and 30 min following the end of the infusion. Template bleeding time was determined before and 30 min after the infusion. In three patients with type 1 VWD the response to DDAVP was studied before treatment and 30 min, 3 h and 4 h thereafter.

VWF concentrates

Very high purity VWF concentrates (LFB, France) were infused to nine VWD patients unresponsive to DDAVP either prior to surgery or to manage bleeding episodes.

Four patients with type 3, two patients with type 2M (type ‘B’), two patients with type 2A and one patient with type 1 ‘platelet-low’ were treated. All patients with type 3 received Innobrand®, which contains both VWF and FVIII; patients with type 2 or type 1 ‘platelet low’ were infused with Facteur Willebrand®, a VWF concentrate which contains only very small amounts of FVIII. VWF concentrates were given at a dose of 60 U VWFRCo/kg body weight (except the patient with type 1 ‘platelet-low’ who received 40 U VWFRCo/kg). Blood was drawn before and 30 min after the end of the infusion and bleeding time was determined before and 30 min after. In one patient with type 3 the response to Innobrand was studied before and 30 min, 3 h and 6 h thereafter.

Laboratory studies

Venous blood was drawn by the use of 19-gauge needles into plastic tubes containing 1/10 final volume of 3.8% sodium citrate. FVIII activity was performed by a one-stage clotting assay based on the aPTT using FVIII-deficient plasma (Diagnostica Stago, France) on the STA automate (Diagnostica Stago, France). VWFAg was measured by ELISA with a commercial kit (Asserachrom VWF from Diagnostica Stago, France). VWFRCo was assayed by aggregometry using a commercially available kit from Behring (Marburg, Germany) which consists of lyophilized platelets and ristocetin A. All results are the mean of two determinations and are expressed as IU/dl of plasma. The Third International Reference Preparation for Factor VIII-related Activities (National Institute for Biological Standards and Control, London) was used as a standard. Multimeric composition of plasma VWF was estimated by sodium dodecyl sulphate gel electrophoresis as previously described (Meyer et al, 1980). RIPA was carried out in platelet-rich plasma on an aggregometer (Thrombo-Agregametre, Regulest, France) with ristocetin (Diagnostica Stago, France) for a minimum of three final concentrations: 0.5, 1.0 and 1.5 mg/ml.

The template bleeding time was determined by a modified Ivy technique using a sterile disposable device (Simplate®, Organon Teknika Corp., Durham, N.C.) according to the instructions of the manufacturer.

The evaluation of the CT using the automate PFA-100 was determined on whole citrated blood (Kundu et al, 1996; Fressinaud et al, 1998). The PFA-100 is a high shear-inducing device that simulates primary haemostasis after injury to a small vessel. Disposable test cartridges contain a reservoir for whole blood and a small capillary surmounted by a collagen-coated membrane with a central aperture. A platelet agonist which is either epinephrine or ADP is present on the membrane. During the test, the blood sample is aspirated at high shear rates through the capillary and comes into contact with the collagen and the agonist. Thus platelets adhere, then aggregate until a platelet plug occludes the aperture. The time required to stop the blood flow and to obtain occlusion of the aperture is defined as the CT. All blood samples were tested with both types of cartridges (collagen/epinephrine and collagen/ADP) as previously described (Fressinaud et al, 1998). The normal range for the CT was calculated as mean ± 2 SD of 96 healthy volunteers (Fressinaud et al, 1998).

RESULTS

PFA-100 and response to desmopressin

Desmopressin was infused in 34 patients with VWD at the time of diagnosis to establish the type of individual response.

Twenty-three patients exhibited all laboratory criteria of type 1 VWD. Before infusion the mean level of VWFAg was 32.5 ± 9.4 IU/dl (range 15–56) and that of VWFRCo 29.1 ± 8.1 IU/dl (range 14–49). The BT was normal (leqslant R: less-than-or-eq, slant 9 min) in 11 patients (47.8%), moderately prolonged (9.5–15 min) in eight (34.8%) and markedly prolonged (> 15 min) in four cases (17.4%) (Fig 1). All patients had a prolonged CT with both types of cartridges. The mean value of CT was 225.2 ± 33.2 s (range 167–250) with collagen–epinephrine and 197.8 ± 49.9 s (range 128–250) with collagen–ADP. The CT was > 250 s with both cartridges in eight patients (Fig 1).

Figure 1.

. Response to desmopressin in 23 patients with type 1 von Willebrand disease. von Willebrand factor ristocetin cofactor activity (VWFRCo), bleeding time (BT) and closure time with the PFA-100 using ADP (CT-ADP) as well as epinephrine (CT-Epi) cartridges were evaluated before and 30 min after infusion of desmopressin in 23 patients with type 1 von Willebrand disease.

The same parameters were evaluated 30 min after the end of the infusion of desmopressin (Fig 1). The mean factor increase over baseline was 2.68 (range 1.39–3.58) for VWFAg and 2.86 (range 1.50–4.14) for VWFRCo. Therefore the mean levels of VWFAg and VWFRCo were 87.7 ± 33.6 and 82.7 ± 32.0 IU/dl respectively. The BT was normalized in all but one patient. The CT was normalized in all patients with both types of cartridges.

In three patients with type 1, response to desmopressin was also studied 3 and 6 h later (Fig 2). The CT was still shortened after 3 h in two patients and returned to basal level in the third. 6 h later the CT was again > 250 s in all patients.

Figure 2.

. Time duration of the response to desmopressin in three patients with type 1 von Willebrand disease. von Willebrand factor ristocetin cofactor activity (VWFRCo), bleeding time (BT) and closure time with the PFA-100 using ADP (CT-ADP) as well as epinephrine (CT-Epi) cartridges were evaluated before, 30 min, 3 h and 6 h after infusion of desmopressin in three patients with type 1 von Willebrand disease (○, patient 1; ▪, patient 2; ●, patient 3).

In five patients with type 2M, including one with ‘type B’, the mean baseline of VWFAg was 36.0 ± 19.4 IU/dl (range 12–53) and that of VWFRCo 13.2 ± 7.6 IU/dl (range < 3 for ‘type B’ to 22). The BT was > 20 min in two patients (one with ‘type B’), markedly prolonged (geqslant R: gt-or-equal, slanted 15 min) in two patients and in the normal range in one patient. The CT was infinite (> 250 s) in all patients with both cartridges (Fig 3). 30 min after the end of the infusion of desmopressin the mean level of VWFAg was 70.6 ± 21.1 IU/dl (range 49–101) and that of VWFRCo was 40.0 ± 32.4 IU/dl (range < 3–90). The BT was normalized in all patients except the one with ‘type B’. The CT was normalized in five patients and unchanged (> 250 s) in the ‘type B’ patient (Fig 3).

Figure 3.

. Response to desmopressin in 11 patients with type 2 von Willebrand disease. von Willebrand factor ristocetin cofactor activity (VWFRCo), bleeding time (BT) and closure time with the PFA-100 using ADP (CT-ADP) as well as epinephrine (CT-Epi) cartridges were evaluated before and 30 min after infusion of desmopressin in 11 variants of von Willebrand disease: three type 2A (♦), two type 2B (□), one type 2B ‘New York’ (▪), four type 2M (○) and one type 2M ‘B’ (●).

In two patients with type 2A the response to desmopressin was excellent, with a mean increase of VWFAg and VWFRCo 3.53 and 3.82 times respectively above the basal levels; the BT (> 20 min) and the CT (> 250 s) were normalized. In the other patient with type 2A the increase of VWFAg and VWFRCo was weaker (1.43 and 1.62 times respectively) and the BT and the CT remained considerably abnormal (Fig 3).

Three patients with type 2B (one with type I New York and two related patients with the classic type 2B) also received desmopressin. All exhibited an increase of VWFRCo levels around 5 times above the basal levels. In the patient with ‘type I New York’, CT was normalized and BT was shortened whereas in the two patients with classic type 2B, the only CT performed with collagen–ADP cartridges was unchanged (Fig 3). The platelet count was within the normal range in the three patients before infusion and slightly decreased only in the two patients with classic type 2B.

PFA-100 and response to VWF concentrates

VWF concentrates were infused in nine patients already known to be unresponsive to DDAVP (Fig 4).

Figure 4.

. Response to VWF concentrates in nine patients with various types of VWD. von Willebrand factor ristocetin cofactor activity (VWFRCo), bleeding time (BT) and closure time with the PFA-100 using ADP (CT-ADP) as well as epinephrine (CT-Epi) cartridges were evaluated before and 30 min after infusion of VWF concentrates in patients with various types of von Willebrand disease: four type 3 (□) who received Innobrand, one type 1 ‘platelet low’ (▪), two type 2A (♦) and two type 2M ‘B’ (●) who received Facteur Willebrand.

Four patients with type 3 VWD had undetectable VWFAg and VWFRCo levels, BT > 20 min and infinite CT. 30 min after the infusion of 60 U VWFRCo/kg Innobrand, VWFAg and VWFRCo levels were normalized with a mean value of 128.8 IU/dl (range 118–138) and 117 IU/dl (range 105–124) respectively. BT was slightly shortened to 15–18 min; CT was unchanged (> 250 s) with both types of cartridges (Fig 4). In one patient BT and CT were also evaluated after 3 and 6 h; BT was still slightly shortened 3 h later and returned to its basal level 6 h later whereas the CT remained > 250 s.

Two patients with type 2M ‘type B’ received the Facteur Willebrand concentrate (60 U VWFRCo/kg). Before infusion, both exhibited undetectable VWFRCo levels and moderately decreased VWFAg levels (32 and 41 IU/dl), BT was > 20 min and CT was infinite. After the concentrate infusion the VWFRCo levels were normalized (88 and 97 IU/dl), BT shortened to 14 and 17 min, and CT was unchanged (> 250 s) (Fig 4).

Two patients with type 2A received the Facteur Willebrand concentrate at the same dosage. The VWFRCo levels were normalized (116 and 125 IU/dl), BT shortened to 15 and 16 min, and CT remained > 250 s (Fig 4).

One patient with type 1 ‘platelet low’ was also infused with the Facteur Willebrand concentrate at a dosage of 40 U VWFRCo/kg. BT was > 20 min, CT-Epi > 250 s and CT-ADP 175 s before infusion; these parameters shortened but remained abnormal 30 min after: 12 min for the BT, 191 s for the CT-Epi and 129 s for the CT-ADP (Fig 4).

DISCUSSION

Given its high prevalence, VWD should be suspected in any patient with a bleeding history. The screening of VWD is considerably improved by the use of the PFA-100 automate which can advantageously replace the BT. In a large study performed in 60 patients with different types of VWD and 96 healthy volunteers, we have previously shown (Fressinaud et al, 1998) that the evaluation of the CT with the PFA-100 is strikingly more reproducible and sensitive, exhibiting a sensitivity > 95% in contrast to 50% for the BT. Moreover, the PFA-100 test provides excellent predictive values > 90% (Fressinaud et al, 1998). Finally, it has the advantage of simplicity and ease of execution, it is not invasive and provides fast results. We confirmed in this new study that the sensitivity of the PFA-100 to detect type 1 VWD is superior with a prolonged CT observed in 100% of cases versus 54.2% for the BT.

It was therefore logical to evaluate the interest of the PFA-100 in the therapeutic monitoring of VWD patients. It is well known that patients having normal levels of functional VWF in storage sites exhibit an effective response to desmopressin, with a normalization of the BT and of the FVIII/VWF levels (Mannucci, 1997). In routine and clinical conditions, however, it is not easy to evaluate the VWF cellular compartment, in particular the platelet VWF content. We have therefore evaluated the CT before and after infusion of desmopressin in 23 patients fulfilling the diagnostic criteria of type 1 VWD. The CT was fully corrected in all patients and this effect appeared to last between 3 and 6 h. The BT was normalized in 10 patients and shortened in one. Therefore we can assume that patients who are good responders to desmopressin have the usual type 1 ‘platelet normal’.

The response to DDAVP was also studied in five patients with type 2M, four with the common missense mutation Arg1374His (Meyer et al, 1997) and one (‘type B’ variant) with a missense mutation Gly1324Ala (Rabinowitz et al, 1992). Desmopressin normalized the CT in the four patients with the common mutation but not in the one with the type B variant. The correction of the CT in the four patients was accompanied by an increase in the plasma VWF levels, similar to that seen in patients with type 1 VWD following DDAVP therapy. In contrast, the ‘type B’ patient had a moderate increase of VWFAg but the VWFRCo level continued to be undetectable. Thus, among patients with the 2M phenotype, i.e. decreased VWFRCo/VWFAg ratio, no loss of high MW multimers, decreased ristocetin — but normal botrocetin — induced binding of VWF to platelets, the response to DDAVP appears to vary according to the genetic defect. Therefore the evaluation of the CT may rapidly and accurately select the good responders.

Among three patients with type 2A VWD two normalized their CT (and VWF levels); in the remaining patient the CT was not shortened and the VWF levels were poorly increased. Again among type 2A patients the measurement of the CT after DDAVP may easily distinguish the two groups of patients (Lyons et al, 1992), differentiating responders (characterized by an increased proteolysis following secretion of fully multimerized VWF) from non-responders (characterized by a defect in secretion of the highly multimerized forms of VWF).

Although the use of desmopressin is controversial in type 2B, three patients received an infusion. One with a mutation responsible for the subtype ‘type I New York’ exhibited an excellent response and no thrombocytopenia. The two other patients with classic type 2B who were related, had a fair increase of VWFRCo level but the CT remained consistently prolonged and thrombocytopenia occurred. In ‘type I New York’ where all multimers are present in plasma, the abnormal VWF has less affinity for the platelet GPIb than the classic 2B VWF; its release after desmopressin thus does not lead to thrombocytopenia. In type 2B, where the abnormal VWF has a high affinity for platelet GPIb, the binding of the released VWF to platelets occurs before its binding to collagen, preventing the platelet anchoring to collagen-bound VWF.

Patients already known to be unresponsive to DDAVP were monitored with VWF concentrates to manage either a surgical procedure or a bleeding episode. Therefore we evaluated the PFA-100 analyser, before and after infusion of Innobrand in four patients with type 3. Although there was a normalization of the plasma levels of VWF, the CTs > 250 s were almost unchanged, and, as is well known (Mannucci et al, 1992), the prolonged BT was hardly modified. Two hypotheses may explain our findings. First, the concentrate infusion does not correct the platelet VWF defect, and many basic and clinical studies have emphasized the key role of platelet VWF in mediating platelet adhesion and aggregation to injured vessels especially at very high shear rates (Gralnick et al, 1986; Castillo et al, 1991, 1997; Fressinaud et al, 1994; Mannucci, 1995). Secondly, no concentrate today exhibits an intact multimeric structure (Mannucci et al, 1994) and the very large VWF multimers (like those contained in storage sites) are known to be the most effective in primary haemostasis. The very high shear conditions of the PFA-100 system probably enhanced these features, making the PFA-100 more sensitive to the platelet VWF and to the presence of large multimers than the BT. Mannucci (1995) established that the correction of defective primary haemostasis (by addition of platelet VWF to plasma VWF) was only necessary to stop mucosal bleeding; this correction was not required to manage soft tissue bleeding nor to prevent surgical bleeding; indeed, none of our type 3 patients exhibited any abnormal bleeding.

One patient diagnosed with type 1 ‘platelet low’, two with type 2M ‘type B’, and two with type 2A who were all unresponsive to DDAVP also received a highly purified VWF concentrate (Facteur Willebrand). In the patient with the subtype ‘platelet low’ both CT and BT were shortened but not normalized. In all the other patients the CT was unchanged and a full correction of the BT was not observed. None of these patients has a normal platelet VWF content as the latter exhibited either a quantitative or a qualitative defect. Again we can assume that the very high shear conditions in the PFA-100 system make it very sensitive to the contribution of platelet VWF and to the ultralarge VWF multimers which are consistently lacking in the concentrates.

Recently the ‘in vitro’ ability of different FVIII/VWF concentrates and cryoprecipitate to shorten the CT of type 3 VWD blood was investigated (Chang et al, 1998). The CT was not well correlated to the VWFRCo levels of concentrates but it was clearly influenced by their content in HMW multimers; not surprisingly, cryoprecipitate provided the best results although full correction of the CT was not reached. In this in vitro study, 100 IU/dl of VWFRCo did not correct the CT and higher concentrations up to 300–400 IU/dl were required. In vivo, however, as mentioned above, a satisfactory clinical outcome is obtained in most situations when plasma VWFRCo levels are around 100 IU/dl.

In conclusion, the PFA-100 system is presently of no use to estimate the efficacy of VWF concentrates in patients who exhibit an abnormality of the platelet VWF content. Conversely, because of its extreme sensitivity to the VWF released from platelets, the evaluation of the CT is a very simple and rapid tool to discriminate between good responders and non-responders to desmopressin.

Acknowledgements

We thank C. Badaud for technical assistance and A. Marville for excellent typing.

Ancillary