Search

We studied four patients who showed aggregation of platelets in platelet-rich plasma at lower concentrations of ristocetin than those required for normal platelet-rich plasma and who demonstrated an increased capacity of the platelets to bind normal von Willebrand factor. The four patients were from two Japanese families. Platelets from one family aggregated spontaneously in vitro, and platelets from both families aggregated upon the addition of normal plasma and cryoprecipitate, in the absence of ristocetin or other agonists. Analysis of the multimeric composition of von Willebrand factor by sodium dodecyl sulfate-agarose gel electrophoresis revealed a decrease in large multimers or a decrease in both large and intermediate multimers in plasma, but normal multimers in platelets. 1-Deamino-[8-D- arginine]-vasopressin caused by an immediate appearance of larger multimers in plasma, followed by the rapid disappearance of these multimers from circulating plasma. Analysis of platelet membrane glycoproteins from the patients showed that there were two distinct bands in the glycoprotein I region; one migrated in a slower region and the other in a faster region than normal glycoprotein Ib. We suggest that the platelet receptor abnormality in these patients is related to this abnormality of glycoprotein Ib.

A case is reported of a 49-year-old woman with a mild bleeding tendency. Her bleeding time, platelet count and size, plasma ristocetin cofactor activity, von Willebrand factor (vWF) antigen, and vWF multimeric pattern are all within normal limits. Spontaneous platelet aggregation is observed when citrated platelet-rich plasma (PRP) is stirred in an aggregometer cuvette. This aggregation is completely is only slightly diminished by an antiglycoprotein (GP) IIb/IIIa or by an anti GPIb monoclonal antibody. The patient's PRP shows increased sensitivity to ristocetin. The distinct feature of this patient, also present in two family members studied, is that platelet aggregation is initiated by purified vWF in the absence of any other agonist. The vWF- induced platelet aggregation is abolished by anti-GPIb and anti- GPIIb/IIIa monoclonal antibodies and by EDTA (5 mmol/L). Apyrase inhibits the second wave of aggregation. Patient's platelets in PRP are four to six times more reactive to asialo vWF-induced platelet aggregation than normal platelets. The amount of radiolabeled vWF bound to platelets in the presence of either low concentration of ristocetin or asialo vWF was increased 30% compared with normal. The patient's platelet GPIb was analyzed by SDS page and immunoblotting and by binding studies with anti-GPIb monoclonal antibodies showed one band with slightly increased migration pattern and a normal number of GPIb molecules. Unlike the previously reported patients with pseudo or platelet-type von Willebrand disease, this patient has normal vWF parameters.

We have analyzed the multimeric structure of factor VIII/von Willebrand factor in plasma by sodium dodecyl sulfate electrophoresis using gels of varying porosity and a discontinuous buffer system. Factor VIII/von Willebrand factor bands were identified by reaction with 125I-labeled affinity-purified antibody and subsequent autoradiography. In 1% agarose gels, normal plasma displayed a series of sharply defined oligomers. However, increasing the agarose concentration to 2.0% or utilizing mixtures of 0.8% agarose--1.75% acrylamide revealed two bands of lesser intensity interposed between the major bands. When the acrylamide concentration in the gels was increased to 2.5%, bands with a faster mobility than IgM and fibronectin were now evident. Type IIA von Willebrand's disease showed not only an absence of the larger multimers but also a relative increase in several of the newly identified bands as compared to type IIB, type I, and normal. These studies suggest that factor VII/von Willebrand factor in IIA von Willebrand's disease is structurally different from that in other forms of the disorder. They also indicate that the multimeric composition of factor VII/von Willebrand factor is more complex than can be explained by simple linear polymerization of a single protomer.

Precipitating antibodies to factor VII/von Willebrand factor can develop in patients with severe homozygous-like von Willebrand's disease following multiple transfusions with blood derivatives. This study of 4 patients treated with cryoprecipitate for 13 different bleeding episodes demonstrates that the occurrence of such antibodies interferes with the management of the disease. The control of mucosal bleeding was poor, whereas more favorable responses were obtained in soft-tissue hemorrhages. These findings probably relate to failure of replacement therapy to shorten the prolonged bleeding time. Immediately after treatment, measurement of plasma factor VIII/von Willebrand factor-related antigen and ristocetin cofactor showed either no increase, or very low values, depending on the pre-infusion antibody titer. Levels of the factor VIII/von Willebrand factor-related procoagulant activity in the circulation were also lower than predicted and usually there was no evidence of the delayed and sustained rise typically observed in uncomplicated von Willebrand's disease. An anamnestic rise in antibody titer appeared 6–15 days after treatment and showed no obvious relationship with the amount of cryoprecipitate infused. Replacement therapy invariably caused severe side effects during, or immediately after, concentrate infusion. The results of in vitro studies support the view that these reactions were due to the appearance of circulating immune complexes.

The effect of intravenous 1-deamino (8-D-arginine)vasopressin (DDAVP) administration on platelet interaction with human artery subendothelium was investigated with flowing blood from five normal individuals and 12 patients with von Willebrand's disease (vWD). Three of the patients were diagnosed as vWD subtype I, four as subtype IIa, and five as subtype IIb. DDAVP administration to normals enhanced platelet adherence, in parallel with increasing plasma levels of factor VIII- related antigen ( FVIIIR :Ag) and ristocetin cofactor activity ( FVIIIR :RCF). Platelet aggregate formation was transiently increased within 90 minutes. Platelet adherence in patient blood before DDAVP infusion was subnormal. In patients with subtype I, administration of DDAVP normalized the bleeding time, enhanced the platelet adherence, and transiently improved the platelet aggregate formation. The platelet adherence was more corrected than would have been expected on the basis of the FVIIIR :Ag and FVIIIR :RCF levels. In patients with subtype IIa, infusion of DDAVP increased the FVIIIR :Ag levels approximately threefold, without affecting the FVIIIR :RCF levels, and in only two of four patients was a transiently enhanced platelet adherence with a corresponding shortening of the bleeding time observed. In patients with subtype IIb, administration of DDAVP increased the FVIIIR :Ag levels about threefold and the FVIIIR :RCF levels five to tenfold, but decreased the platelet adherence significantly. The bleeding time values were not normalized. A close association between the bleeding time values and corresponding platelet adherence values before and after DDAVP infusion was observed. Normalization of the bleeding time was paralleled with normalization of platelet adherence. We conclude that DDAVP improves the primary hemostasis by causing enhanced FVIII- vWF-mediated platelet adherence. DDAVP has little or no effect on the bleeding time in patients with subtype IIa and subtype IIb, because the platelet adherence is not normalized.

The multimeric structure of platelet factor VIII/von Willebrand factor (FVIII/vWF) in cell extracts and in collagen and thrombin releasates has been analyzed by SDS polyacrylamide gel electrophoresis followed by detection with 125I-anti-FVIII/vWF. Platelets contained larger multimers than those normally present in plasma. When secreted FVIII/vWF was analyzed, all platelets. In contrast, in thrombin releasates the larger multimers were lost in a manner dependent on divalent cations, time, and thrombin dose. This loss could not be accounted for by modification of FVIII/vWF by thrombin or platelet enzymes since no effect of thrombin on the multimeric structure of FVIII/vWF in the absence of platelets or in the presence of platelet lysates was observed. Large multimers of 125I-labeled purified FVIII/vWF underwent divalent cation-dependent association with platelets in the presence of thrombin, indicating that the loss of FVIII/vWF from thrombin releasates was due to reassociation with the platelet. These studies show a structural difference between platelet and plasma FVIII/vWF that suggests a specific role for platelet FVIII/vWF in hemostasis.

In order to evaluate the role of proteolysis in acquired von Willebrand's disease (vWD) associated with the myeloproliferative syndrome, we have determined the relative quantity of von Willebrand factor (vWF) fragments as compared with the intact 225 kDa subunit in four patients. The plasma vWF of each individual lacked large multimers; each had a prolonged bleeding time; and both platelet and leukocyte counts were elevated. Plasma was obtained from blood drawn into 1 mmol/L leupeptin, 6 mmol/L N-ethylmaleimide, and 5 mmol/L EDTA to prevent in vitro proteolysis. vWF was isolated from plasma by immunoadsorbent chromatography, reduced, subjected to SDS-5% polyacrylamide gel electrophoresis, and immunoblotted with a mixture of 55 anti-vWF monoclonal antibodies. In three patients with essential thrombocytosis (ET) the 176 and 140 kDa fragments were increased in proportion to the intact 225 kDa subunit indicating increased proteolysis. Treatment of one ET patient with CCNU (Lomustine) decreased the platelet count and, to a lesser extent, the white blood cell count. This was associated with a correction of the bleeding time, a partial correction of the multimeric abnormality, and a lessening of vWF cleavage. In a patient with polycythemia rubra vera (PRV) the proportion of the 176 kDa fragment was increased to the upper limit of normal but there was no definite evidence of increased proteolysis. These studies provide evidence that proteolysis plays a role in the acquired von Willebrand's disease associated with the myeloproliferative syndrome. However, other mechanisms must also be considered.

The absence of large von Willebrand factor (vWF) multimers from plasma is a characteristic of Type IIA von Willebrand's disease (vWD) and is thought to contribute to the clinical expression of this disorder. Recently, three IIA patients have been reported in whom intermediate and large multimers could be restored if blood were collected in 5 mm EDTA, 6 mmol/L N-ethylmaleimide, and 1 mmol/L leupeptin. This suggested that absence of large multimers resulted from in vitro proteolysis. We have now collected blood from ten Type IIA vWD patients in these inhibitors but were not able to detect large multimers in the plasma of any of them. In addition, intermediate-sized multimers were reduced or completely absent in all. The inclusion of inhibitors in the citrate anticoagulant, as compared to citrate alone, was found to increase the relative proportion of intermediate multimers in some patients but had no effect in others, and in none did it restore large multimers to plasma. The results with platelet vWF were more varied. Four patients showed an absence or decrease of large multimers, whereas in seven patients large multimers were present. When compared with citrate anticoagulant alone, the inclusion of inhibitors in the anticoagulant had little or no effect on the platelet multimeric pattern. 1-Deamino-8- D-Arginine Vasopressin (DDAVP) was administered to six patients from five families. Two patients from one family showed complete correction and a third patient showed almost complete correction of her bleeding time. Two patients showed minimal correction and one showed no detectable correction. An increase in multimer size after DDAVP tended to be associated with correction of the bleeding time. However, in no case did the largest multimers appear in plasma even in patients with complete bleeding time correction. The presence or absence of inhibitors in the anticoagulant had little or no effect on the multimeric pattern after DDAVP. These results indicate that Type IIA vWD is a heterogeneous disorder in which absence of largest and intermediate multimers is an in vivo phenomenon.

A patient with a lifelong bleeding disorder was diagnosed as having Type II von Willebrand disease. The larger multimers of von Willebrand factor were absent from her plasma but present in platelets. A high- resolution electrophoretic technique was used to study the complex structure of individual von Willebrand factor multimers. In normal plasma, each multimer could be resolved into five bands: a more intense central one and four less intense, two moving faster and two slower than the central band. In normal platelets, each multimer could also be resolved into five bands. The central one had a mobility similar to that in plasma, whereas the four satellite bands had a mobility that differed from that of the corresponding plasma bands. In the patient, platelet von Willebrand factor antigen content and ristocetin cofactor activity were normal, and von Willebrand factor showed the same structure of individual multimers as seen in normal platelets. On the other hand, plasma von Willebrand factor antigen and ristocetin cofactor activity were decreased, and the structure of individual von Willebrand factor multimers was different from that of normal plasma and similar to that seen in normal and patient's platelets. After infusion of 1-deamino-8-D-arginine vasopressin, the largest von Willebrand factor multimers, as well as new satellite bands with a mobility similar to those in normal plasma, appeared in the patient plasma, and the levels of von Willebrand factor antigen and ristocetin cofactor activity became normal. Yet no relevant change in the prolonged bleeding time was observed. This new variant of von Willebrand disease, therefore, is characterized by the presence of a dysfunctional von Willebrand factor molecule that exhibits unique structural abnormalities in plasma but appears to be normal in platelets. The designation of Type IIF is proposed for this type of von Willebrand disease in accordance with the terminology that has been previously used.

The infusion of 1-deamino-(8-D-arginine)-vasopressin (DDAVP) causes not only an elevation in factor VIII-related antigen (FVIIIR:Ag), but also a marked elevation of plasma von Willebrand antigen II (vWAgII). vWAgII reaches a peak concentration at 60 min and is elevated 3–8-fold over basal levels in normal individuals and individuals with type I, IIA, and IIB von Willebrand's disease. As the mechanism of hemostatic alteration brought about by DDAVP might be due to release of endothelial cell proteins, endothelial cell cultures were performed. The cultures demonstrated synthesis and secretion of vWAgII, as evidenced by the incorporation of 35S-methionine into the vWAgII molecule. Thus, vWAgII, like FVIIIR:Ag, is an endothelial cell protein.

Botrocetin, a component of Bothrops jararaca venom, induces von Willebrand factor (vWF)-dependent platelet agglutination and has been proposed as an alternative agent to ristocetin for evaluating vWF function. However, important differences between the vWF-platelet interactions induced by these two agents have suggested that different regions of vWF and the platelet may be involved in the interactions induced by the two agonists. We have recently demonstrated that binding of vWF to the platelet glycoprotein (GP) Ib receptor, either induced by ristocetin or as occurs spontaneously with asialo-vWF or vWF from IIb von Willebrand disease, is mediated by a domain residing on a 52/48- kilodalton (kD) tryptic fragment of vWF. This fragment extends from amino acid residue Val (449) to Lys (728). We have now found that this 52/48-kD fragment blocks botrocetin-induced binding of vWF to platelets and completely inhibits botrocetin-induced platelet agglutination. These results provide evidence that the vWF domain-mediating, botrocetin-induced platelet agglutination lies within the region delimited by this fragment and is therefore close to or identical with that which mediates ristocetin-induced binding and spontaneous binding of vWF to platelet GPIb. Anti-GPIb monoclonal antibodies also blocked agglutination, which showed that botrocetin, like ristocetin, induces binding of vWF to the GPIb receptor.

Progress has occurred in the past several years in the understanding of the structure and function of von Willebrand factor (vWF). This multimeric glycoprotein exhibits a dual role, that of mediating platelet adhesion and aggregation onto thrombogenic surfaces, and that of functioning as carrier in plasma for the factor VIII procoagulant protein. New insights into the nature of the several functional domains of vWF have led to the identification of the regions of the molecule that interact with factor VIII, heparin, the glycoprotein lb of platelets, and collagen. Alterations of vWF are the cause of von Willebrand disease (vWD), a congenital bleeding disorder. In the majority of patients, the plasma levels of vWF are decreased, but there is no demonstrable structural or functional alteration of the protein. In other patients, however, the structure of vWF is abnormal. This review summarizes the current knowledge on vWF and vWD.

Platelet activation altered the binding of three monoclonal antibodies (monovalent Fab' fragment) directed against the glycoprotein (GP) IIb/IIIa complex. An increased binding of two- to threefold occurred after stimulation with thrombin or phorbol myristate acetate (PMA), with slight but significant increase in the dissociation constants (Kd) of two antibodies (LJ-CP8 and LJ-P9). In contrast, no statistically significant changes were observed with ADP-stimulated platelets. The increased binding of LJ-CP3, but not of the other two antibodies, to activated platelets decreased by 30% to 40% in the presence of EDTA at 22 to 25 degrees C. Platelets stimulated by thrombin or PMA bound more fibrinogen than did those stimulated by ADP, and significant differences in the extent but not in the affinity of fibrinogen binding were observed with various platelet agonists. When the pool of GP IIb/IIIa molecules exposed on the surface of unstimulated platelets was reacted with the monoclonal antibody LJ-CP3 to block ADP-induced fibrinogen binding and platelet aggregation, stimulation with thrombin or PMA still induced substantial binding of antibody and fibrinogen, and aggregation ensued. Therefore, platelets exposed to “strong” agonists exhibit an increased number of surface-oriented epitopes associated with GP IIb/IIIa. The GP IIb/IIIa molecules bearing these newly exposed epitopes are functional in that they can bind fibrinogen and mediate platelet aggregation.

The present studies demonstrate that platelets from patients with platelet-type von Willebrand disease show specific and saturable binding of asialo von Willebrand factor (AS-vWF) under conditions where such binding is not observed with normal platelets. Although specific binding of 125I-AS-vWF to formalin-fixed normal platelets could not be demonstrated, specific binding to fixed patient platelets was seen with an apparent Kd of 1.3 micrograms/mL and specific maximally bound ligand of 0.40 micrograms/10(8) platelets. Preincubation of patient platelets with the antiglycoprotein Ib (anti-GPIb) monoclonal antibody AS-2 reduced total binding close to the level of computer-estimated nonspecific binding. In contrast, binding was not reduced by preincubation with anti-GPIIb/IIIa monoclonal antibody or with 5 mmol/L EDTA. Under stirring conditions, the binding of AS-vWF to fixed patient platelets was accompanied by a strong agglutination response. AS-vWF- induced agglutination was similarly observed in patient but not normal platelet-rich plasma (PRP) in the presence of 5 mmol/L EDTA. In the absence of EDTA, AS-vWF produced a full aggregation response in patient PRP at concentrations as low as 0.1 microgram/mL in contrast to the 2 to 20 micrograms/mL required by normal PRP. Both thromboxane B2 formation and adenosine triphosphate secretion showed an AS-vWF concentration dependence paralleling the aggregation responses. These studies show that a major difference in the platelets from patients with platelet-type von Willebrand disease is the presence of an exposed, high-affinity binding site associated with GPIb that recognizes AS-vWF.

The infusion of 1-deamino-8-D-arginine vasopressin (DDAVP) in normal individuals is followed by an increase in factor VIII/von Willebrand factor (vWF) in plasma, by an increase in intensity of all sizes of multimers, and by the appearance of larger multimers of vWF than those seen in the resting state. Since the larger multimers are rapidly cleared and proteolysis is known to cause disaggregation of large multimers, we evaluated the degree of vWF proteolysis after DDAVP administration. DDAVP was infused into eight normal adult volunteers, and the relative proportions of the intact 225 kilodalton (kDa) subunit and the 189, 176, and 140 kDa vWF fragments were compared before and at different times after DDAVP infusion. The relative proportion of the 176 kDa fragment was increased, whereas that of the other species was decreased, thereby indicating that proteolytic fragmentation had occurred. However, plasmin did not appear to be responsible because the vWF fragments characteristically produced by this enzyme could not be detected. Concomitant analysis of vWF multimeric structure showed that these changes were accompanied by an increase in the relative proportion of the satellite bands, which suggests that they were proteolytically generated. Proteolysis may explain, at least in part, rapid clearance of larger vWF multimers released by DDAVP.

In Type I von Willebrand disease, the whole series of von Willebrand factor (vWF) multimers is present in plasma, but all are decreased in quantity. No structural abnormality of individual multimers has been demonstrated so far in these patients. We now describe five individuals, from two unrelated families, who had this form of the disease and in whom the complex banding pattern of each vWF multimer was markedly abnormal. Inheritance was autosomal dominant and the clinical expression was mild. A bleeding history was elicited in three of the patients and included recurrent epistaxis, menometrorrhagia, and bleeding following tooth extraction. Replacement therapy had never been required. Although vWF levels in plasma were within the normal range in all of them, the ristocetin cofactor activity was decreased in four, and the bleeding time was prolonged in three. Analysis of vWF multimeric structure by agarose gel electrophoresis, including a newly developed high-resolution technique, demonstrated that the main band of each multimer was present, but a second, well-defined band always seen in normal individuals was missing in the patients. Two additional bands had altered mobility and were less well defined than in normal subjects, and a fifth, less intense band was also undetectable in the patients. Treatment with 1-deamino-8-D-arginine vasopressin (DDAVP) was assessed in two patients. It caused the circulating levels of vWF to increase and correct the bleeding time, but did not alter the structural abnormality. This study describes, therefore, a new variant form of Type I von Willebrand disease with aberrant structure of individual repeating multimers and an associated functional abnormality of vWF. In keeping with previously accepted terminology, the designation of Type IC von Willebrand disease has been adopted for this new variant.

Type IIB von Willebrand disease is characterized by enhanced ristocetin- induced platelet aggregation and absence of large von Willebrand factor multimers from plasma. An alteration of the von Willebrand factor molecule resulting in increased reactivity with platelets appears to be the basis for these abnormalities. We have now identified a new variant of type IIB von Willebrand disease in a family in which the four affected members also have chronic thrombocytopenia, in vivo platelet aggregate formation, and spontaneous platelet aggregation in vitro. In spite of repeatedly prolonged bleeding times and persistent thrombocytopenia, their bleeding diathesis is only moderate.

Human bone marrow was separated on continuous Percoll density gradients to analyze the distribution of cells of the megakaryocytic lineage. Megakaryocytes were identified by indirect immunofluorescence using a monoclonal antibody (LJP4) specific to the glycoprotein IIb/IIIa (GPIIb/IIIa) complex of platelets. Neither endothelial cells nor monocytes expressed the epitope identified by this antibody. Simultaneous measurement of size and ploidy were made on 2,359 GPIIb/IIIa-positive cells. The smallest cells were located in the most dense fractions where 81% of all 2N and 66% of 4N cells were found at densities greater than or equal to 1.050 g/mL. The largest cells were detected in the least dense fractions, with 70% of 16N, 78% of 32N, and 100% of 64N cells found at densities less than 1.050 g/mL. Ninety-four percent of all GPIIb/IIIa-positive cells less than 14 micron in diameter were found at densities greater than 1.050 g/mL. An exception to this inverse relationship was observed in the uppermost gradient fractions where an anomalous distribution of size and ploidy was found. Megakaryocytic viability was identified as being greater than 90% in all fractions. The data show that megakaryocytic differentiation as assessed by size and ploidy varies inversely with Percoll density. Separation of marrow on continuous Percoll gradients may be a useful method to separate megakaryocytes at different stages of differentiation.

The abnormal multimeric composition of plasma von Willebrand factor in type IIB von Willebrand's disease is transiently corrected after infusion of 1-deamino-[8-D-arginine]-vasopressin. However, the larger multimers released into the circulation disappear more rapidly in these patients than in type I von Willebrand's disease or normals. We demonstrate that the larger multimers of normal von Willebrand factor transfused into a type IIB patient are cleared from the circulation more slowly than multimers of similar size endogenously released from tissue stores. The rate of disappearance of large von Willebrand factor multimers after infusion of cryoprecipitate is similar in IIB, IIA, and severe homozygous-like von Willebrand's disease. Platelets from the IIB patient exhibited normal ristocetin-induced binding of normal von Willebrand factor. However, like normal platelets, they bound IIB von Willebrand factor at lower ristocetin concentrations than required for normal von Willebrand factor. These findings provide evidence that absence of the larger multimers from IIB plasma is related to a molecular abnormality of von Willebrand factor rather than to enhanced affinity of abnormal tissue or cellular binding sites, as is the case in the recently described “pseud” von Willebrand's disease and “platelet-type” von Willebrand's disease.

An analytical technique has been developed for the isoelectric focusing (IEF) of plasma von Willebrand factor (vWF) in agarose gels containing urea. Under these conditions, vWF freely enters the gel and focuses without artifact. The focused vWF is visualized by staining fixed gels with 125I-labeled affinity-purified heterologous antibody. Utilizing a pH gradient of 5.0–6.5, normal vWF in plasma or purified preparations focuses into at least three bands with apparent isoelectric points (pI) between pH 5.7 and 5.9. A reproducible difference in the IEF pattern of vWF has been established between normal plasmas and those of individuals with variant von Willebrand's disease (vWd) type IIA and type IIB. In type IIA, vWF has a distinctly lower pI than normal. This difference may be related to the presence of smaller vWF multimers in IIA plasma because forms of vWF of corresponding size contained in normal cryoprecipitate supernatant have a similar pI. Type IIB von Willebrand factor has a pI intermediate between normal and IIA. Neuraminidase treatment of plasma samples before IEF results in an increase in pI in normal, type IIA, and type IIB vWF. The data suggest that none of the 16 type IIA and 9 IIB plasmas studied here contain significantly decreased amounts of sialic acid.

We have studied the modifications in the multimeric composition of plasma factor VIII/von Willebrand factor and the bleeding time response following administration of 1-Deamino-[8-D-arginine]-Vasopressin (DDAVP) to patients with different subtypes of von Willebrand's disease. In type I, all multimers were present in plasma in the resting state, though they were decreased in concentration. Administration of DDAVP resulted in an increased concentration of these forms as well as the appearance of larger forms than were previously present. There was concomitant correction of the bleeding time. In type IIA, large multimers were absent in the resting state, and although DDAVP induced an average threefold increase in the plasma concentration of factor VIII/von Willebrand factor, the larger multimers did not appear and the bleeding time, although shortened, was not corrected. In contrast, the larger multimers that were also absent from type IIB plasma in the resting state rapidly appeared following DDAVP administration. However, their appearance was transitory and the bleeding time, as in IIA patients, was shortened but not corrected. The characteristic multimeric composition of platelet factor VIII/von Willebrand factor in given subtypes predicted the alteration in plasma factor VIII/von Willebrand factor induced by DDAVP. These studies provide evidence that the different subtypes of von Willebrand's disease represent distinct abnormalities of factor VIII/von Willebrand factor. They also suggest that complete hemostatic correction following DDAVP can be routinely expected only in type I von Willebrand's disease, and only if factor VIII/von Willebrand factor can be raised to normal levels.