T H E V O N WILLEBRAND FACTOR ° FRANCESCO RODEGHIERO, GIANCARLO CASTAMAN

Divisione di Ematotogia, Centro Regionate per lo Studio delle Malattie Emorragiche e Trombotiche, Ospedale San Bortolo, Vicenza

The von Willebrand factor (vWf) is a large nonenzymatic multimeric and multivalent adhesive protein synthesized by megakaryocytes and endothelial cells 9,28. It is stored in platelet e-granules and in Weibel-Palade bodies o f endothelial cells, and it can be secreted from endothelial cell into circulating plasma or abluminally, in which case the factor becomes part of the subendothelial matrix 19,32. It participates in hemostasis by mediating the adhesion of platelets to exposed subendothelium and by promoting the formation of platelet plug at the site of vascular injuries. Moreover, vWf circulates as a noncovalent complex with coagulant factor VIII (FVlII:C) and stabilizes this labile protein in circulation ~0,36. Yon Willebrand factor works through several molecular domains which bind to cell receptors (particularly Gp Ib and Gp I I b / I I I a on platelet surface), circulating proteins and insoluble components of the subendothelium 7. The formation of the initial monolayer of platelets covering injured endothelial surfaces is particularly believed to involve a bridging function of vWf which binds to the subendothelium (collagen and heparin-like substances) and, on the other side, to Gp Ib 7,2s. This interaction results in platelet activation and receptors Gp I I b / I I I a come also into play and recruit other platelets to the growing thrombus through a platelet/platelet interaction mediated by the bridging of fibrinogen and vWf between these receptors 25. The multivalency of vWf depends on its muldmeric structure since all the molecular domains are present in the constitutive repeating subunit (mol. wt. 2.7 X 105 daltons) 9. Two of these suhunits form a dimer (protomer) by disulfide bonding at the carboxyl-termini (tail-to-tail). Protomers are subsequently joined by covalent disulfide bonds in the amino-terminal regions (head-toKey-words: Coagulation assay; Epidemiology; Standardization; von Willebrand disease; von Willebrand factor. • Presented at the '2nd International Symposium on Standardization and Quality Control of Coagulation Tests: Implications for the Clinical Laboratory', Rome, September 28-29, 1989. Res. Clin. Lab. 20, 143-153, 1990.

143

THE VON v~Ar[lIFBRAND FACTOR

head) to yield multimers up to a molecular weight of 10-20 × 106 daltons. The multivalency of larger multimers may result in greater binding affinity and explains the greater efficacy of the larger multimers in hemostasis. The quantitative or qualitative, acquired or congenital deficiency of vWf leads to an impairment of primary hemostasis and to a deficiency of FVIII:C in plasma. Patients suffer from an hemorrhagic diathesis (yon Willebrand disease, vWd) which presents a wide spectrum of severity ranging from almost asymptomatic cases to cases with a severe bleeding disorder. 1.

LABORATORY INVESTIGATION OF

vWf

(EPIDEMIOLOGICAL APPROACH)

In this review we shall focus on the quantitative determinations of vWf in plasma. The role of these measurements in screening and diagnosis will be dealt with in some detail. In the first part an epidemiotogical approach will be used, whereas in the second part the standardization and the technical performance of these assays will be discussed. At variance with other hemostatic proteins, no single test is available which entirely reflects the various functions of vWf and thus several investigations are required to characterize the defect e n c o u n t e r e d in the individual patient. Table 1 shows the principal tests employed in the study o f vWf and their significance. 1.1 Quantitativedetermination of vWf Because of the heterogeneity of vWd and the variability of laboratory pattern in the same subject over time 1,12, a set of tests including bleeding time and FVIII:C assay in addition to vWf:Ag and ristocetin cofactor (RiCof) is usually employed. Especially in borderline cases, whose prevalence in population has been estimated to be exceptionally high reaching 1-2% 2~, not all tests result invariably abnormal and the diagnosis can be made when at least some o f the symptomatic members of an hemorrhagic family (with symptoms transmitted in an autosomal dominant pattern) present with an abnormality of one or more o f these tests. Prolongation o f bleeding time is absolutely not specific and is interpreted only in the context o f the above mentioned tests. The concentration of FVIII:C reflects only indirectly the level o f its circulating carrier, that is vWf, and the usefulness of FVIII:C assay as an indirect measurement of reduced vWf is very limited since the relationship between the levels o f the two proteins is not strict. However, FVIII:C level determination is helpful in evaluating the clinical severity of the disorder and its assay should always be performed in patients with vWd. The direct quantitative determination o f v W f involves essentially two different approaches. The first is based on the interaction, in presence of ristocetin (or botrocetin), between vWf and Gp Ib on the platelet membrane (RiCof). The second is based o n the interaction between vWf and polyclonal or monoclonal antibodies (vWf:Ag) as revealed by several methods, e.g., electroimmunoassay (EIA), immunoradiometric assay (IRMA), enzyme-linked immunosorbent assay (ELISA). The RiCof activity depends not only on the amount of the factor but also on its ability to bring about this interaction, larger multimers being proportionally more active. This fact is reflected by the disproportionally lower content of RiCof in comparison to that of vWf:Ag in variant 144

F. RODEGHIERO, G. CASTAMAN

test

pathophysiological significance

diagnostic significance

fistocetin cofactor

vWf-Gp Ib interaction as mediated by ristocetin in vitro (ristocetin at fixed concentration, normal platelets, patient plasma)

'fiJnctional' test; most sensitive screening test

immunological assay with polyclonal antibody (vWfiAg)

antigen concentration

correlates with RiCof in types I and III, less reduced in type II

ristocetin-induced platelet aggregation (PUPA)

threshold ristocetin concentration inducing patient ptatelet-rich plasma aggregation

reduced threshold in some vWd variants (type IIb and type I 'New York')

botrocetin cofactor

analogous to RiCof

no definite advantage in comparison to RiCof

bidimensional crossed-immunoelectrophoresis

lack of larger multimers

fast moving peak in type II

multimeric analysis

multimeric composition and inner structure

classification of variant types

subunit composition

altered proteotysis or abnormal subunit structure

as above

binding of vWf to collagen

vWf-collagen interaction

correlates with RiCof

bleeding time

platelet-vessel wall vWf-mediated interaction

not specific; correlation with intraplatetet vWf?.; correlation with clinical symptoms?

FVIII:C level

FVIII/vWf interaction

not specific, but useful for patient management

analysis of intraplatelet vWf

reflects endothelial stores?

functional characterization; useful to predict responsiveness to desmopressin

(PiCot')

Tab. 1 - Laboratory methods for investigating the von Willebrand factor.

type II vWd, lacking high and intermediate molecular weight multimers. Thus, RiCof is improperly considered a 'functional test' even if it does not entirely reflect the 'true' activity of the vWf, since it does not explore the interactions between vWf and Gp I I b / I I I a on platelet surface or the subendothelial substances like collagen, fibronectin or heparin-like glycosaminoglycans. Furthermore, this relationship between molecular weight o f multimers and its RiCof activity is not always tenable, as for example in vWd "Vicenza', where supranormal multimers circulate in plasma and vWf:Ag is invariably higher than RiCofaS. Recently, an ELISA test for the binding of vWf to collagen has been proposed. It correlates with RiCof in normals and patients with vWd and has been proposed as an easier alternative to RiCof; however, its usefulness in clinical practice has to be established 3. The diagnostic and clinical usefulness of these tests depends strictly on a proper definition of their normal range and on the knowledge of their predictive value. 145

T H E VON WILLEBRAND FACTOR

1.2 Definition of normal range of vWf'Ag and RiCof At variance with the hemostatic disorders transmitted in a recessive manner (like factor XIII deficiency or hemophilia), in which only homozygous (or hemizygous) members are affected and the pertinent biological activity is almost severely reduced, in vWd (with the exception of type III) there is an overlapping between the lower limit o f normal population and the upper limit of the patient (heterozygous) population. This makes mandatory that normal range is defined as precisely as possible. As it is customary, the specificity (that is how many of the normal subjects have a normal value) is set at 95% (considering upper and lower limits) or at 97.5% (considering only the lower normal limit as in our case, since we are not interested, in this setting, in subjects with higher than normal vWf). This limit should be estimated in a group of reference subjects thought to represent the patient population. In order to make that this limit calculated from the reference population is really representative o f the population that will be investigated, two aspects are essential: 1. the reference population should be really representative of the patient population; 2. the statistical approach should be properly chosen. Pertinent to point 1. (like has to be compared with like) is the analysis of the influence of physiological parameters on the levels of these tests. Without going into details, we can summarize this influence as follows ~,2s,24: a. no influence of sex; b. weak positive relationship with increasing age (in adults group), especially for group 0 and only for vWf:Ag (not requiring an age adjustment for normal range); c. for children (aging 11-14) a separate normal range should be established, since significantly higher values are found; d. for subjects aging 0-11 and 14-20 no data are available; e. subjects of group 0 have lower vWf:Ag and RiCof; f no influence o f Rh phenotype and secretory status. Pertinent to point 2. is the suggestion to use a nonparametric method since the reference individual values of vWf do not fit a normal distribution, and to

vWf:RiCof level (IU/dl) subjects

children

adults

blood group

n-"

mean (range)

2.5 percentile (90% CI)*

97.5 percentile (90% CI)*

0

510

w~.7 (37-178)

61.0 (56-66)

149.8 (146-154)

non-0

656

115.8 (44-230)

77.0 (76-81)

164.0 (156-176)

0

121

88.7 (46-123)

51.8 (46-55)

118.3 (117-123)

non-0

168

110.6 (70-203)

75.1 (7~81)

167.6 (145-203)

T a b . 2 - D i s t r i b u t i o n o f v W f : R i C o f levels in a r e f e r e n c e p o p u l a t i o n o f c h i l d r e n a n d adults. (* 90% confutence intervals around the percentile).

146

F. RODEGHIERO, G. CASTAMAN

vWf:Ag level (IU/dl) subjects

children

adults

blood group

n -~

mean (range)

2.5 percentile (90% CI)*

97.5 percentile (90% CI)*

0

488

134.t (38-369)

57 (53-61)

255 (247-287)

no n-0

633

162.7 (37-418)

71 (66-74)

302 (282-329)

0

129

84.7 (30-216)

36 (30-43)

175 (138-216)

non-0

156

122.8 (41-254)

62 (41-68)

226 (210-254)

Tab. 3 - Distribution of vWf:Ag levels in a reference population of children and adults. (* 90%confidenceintervalsaroundthepercentile).

use a large reference population to reduce the imprecision of range definitions. Tables 2 and 3 show the normal ranges established in our laboratory for the appropriate group of patients, by taking into consideration the above mendoned points. 1.3 Sensitivity and predictive value of RiCof and vWf'Ag Sensitivity in an epidemiological sense means the percentage o f subjects with the disorder (patients) detected by the test or, in other terms, 'how many of the patients have a positive result'. There is a lack of information regarding the absolute sensitivity of vWf:Ag and RiCof in the diagnosis of vWd. Furthermore, since these tests are used for diagnosis and no independent substitutive tests are presently available, some circularity is unavoidable and we should await for independent tests, like genedc markers, to answer this problem. In the only available study o f MILT~R et al.17.1s this problem has been partially avoided by analyzing in two large families with type I vWd how many of the symptomatic members, thus presumably affected by vWd, would have been detected by either test. These authors showed that only about 50% of symptomatic members were detected by RiCof and 29% by vWf:Ag (no separate ranges for 0 or non-0 group were used). The relative sensitivity of these two tests has been the subject of limited investigations. The higher relative sensitivity o f RiCof was already known from the data of the Italian Working Group study 12. Recently, in our laboratory we have • demonstrated in a large epidemiological study that the relative sensitivity of vWf:Ag in comparison to RiCof was 64% ~4. Assuming an absolute sensitivity of RiCof of 50%, vWf:Ag would be 32%, a value very similar to that shown by Mjr T~R et al. 17,18. Circularity was avoided since the two tests, coupled to clinical symptoms, were used as independent criteria for diagnosis, but we cannot gain from our data any insight into the absolute sensitivity o f these tests. Assuming an absolute sensitivity o f RiCof of 50%, a relative sensitivity of vWf:Ag of 64%, a specificity for both tests of 97.5%, and a prevalence o f vWd of 2% in general population and of 30% in hemorrhagic population (the prevalence was estimated in hemorrhagic patients referred to our laboratory), one can calculate the parameters shown in tab. 4. 147

THE VON WILLEBRAND FACTOR

One can observe that only about 10% of hemorrhagic patients would be falsely classified as having vWd using RiCof, whereas in fact has a different hemorrhagic disorder. In the general population the misclassification rate would be about 70%, and thus RiCof cannot be usefully applied as a mass screening test. Similar values are obtained using vWf:Ag. However, this does not mean that this more simple test can replace RiCof in diagnosis of vWd, since its sensitivity is significantly lower and RiCof should remain, whenever possible, the preferred screening and diagnostic test. Recently, the quantitation of intraplatelet vWf, either as RiCof or vWf:Ag, and the analysis of its multimeric composition have attracted an increasing interest for a more complete characterization of subtypes of type I vWd 8,14,26 This is more than of academic interest, since it has been demonstrated that the normalization of bleeding time after 1-desamino-8-D-arginine vasopressin (DDAVP) requires normal concentration and function of intraplatelet vWf, which is thought to reflect the vWf stored in the endothelial cells14.2L It is interesting to note that, on the basis of preliminary results from our laboratory, blood group does not seem to influence the intraplatelet content o f vWf.

2.

ANALYTICAL VARIABLES AND TECHNICAL ASPECTS IN THE PERFORMANCE

OF RiCof AND vWf:Ag ASSAYS 2.1 RiCofassay The assay is based on the observation that a log-log relationship exists between the extent of ristocetin-induced aggregation of washed normal platelets and the "concentration of normal plasma35; that is a plasma factor, subsequently called ristocetin cofactor, can be measured in this way. Subsequently, MEYER et a1.16 showed that a similar relationship existed between the amount o f aggregation, measured as the slope of the steepest part of the agglutination trace, and the ristocetin cofactor activity. This way to quantitate platelet aggregation remains the most used approach and also in our hands it has proved to be the most precise and sensitive method. A great improvement was provided by MACFARLANEet al. 1., who showed that formalin-fixed platelets could be used

vWfi~Cof

vWfiAg

hemorrhagic population

general population

hemorrhagic population

general population

predictive value of a positive result (PVe %)

89.6

28.9

84.6

20.9

predictive value of a negative result (PV, %)

82.0

98.9

76.9

98.5

diagnostic efficiency (DE, %)

83.3

96.6

77.8

96.1

Tab. 4 - Predictive values and diagnostic efficiencyof vWf:RiCofand vWfiAgin hemorrhagic a n d g e n e r a l p o p u l a t i o n . (The follozaing assumptions were used for calculation: specificity 97.5%; sensitivity 50%for RiCof and 320 for vWf'Ag; vWd prevalence in general population 2%, in hemorrhagic patients 30%. The following formulas were used for calculation: PVp = TP ; PV~ = TN ; DE = TP + T N ; TP + FP 23V + FN TP + TW + FP + F N where TP are true positives, FP false positives, T N true negatives and F N false negatives).

148

F. RODEGHIERO, G. CASTAMAN

- formalin- o r glutaraldehyde-flxed platelets - r e c o r d rate o f aggregation - test sample at three different dilutions - replicate calibration curve (at b e g i n n i n g a n d at the e n d o f the assay) - duration o f the assay not longer than 2h - use statistical a p p r o a c h in calculating results (parallel line bioassay) - control p l a s m a with low a n d high vWf content tested in each assay - calibrate y o u r local standard against IRP (using vWf-deficient p l a s m a for dilution in dose-response curves if lyopbilized standard is involved ~) - a u t o m a t i o n advisable

Tab. 5 - Technical suggestions for the performance of vWf:RiCofassay by an aggregometric method.

with the advantage that platelets could 'usually be stored at 4°C for several weeks'. We usually observe a loss of responsiTeness after 2-3 weeks o f storage. In our laboratory the test sample is tested at three different dilutions and its ratio to the standard curve (internal standard, usually normal lyophilized plasma diluted 1:1 to 1:32) is calculated using parallel line bioassay methodology 21. It is of paramount importance that the internal standard is calibrated in international units (IU) against the InternationalReference Preparation (IRP) 2. With the most reactive platelet preparation, the level that can be measured is about 3 IU, whereas usually values below 6 IU are not detected. Very few data are available concerning the precision o f this assay. During the first international multicentre study undertaken for the calibration of a reference preparation for factor VIII/vWf-related activities in plasma s, the within-laboratory precision (measured as geometric coefficient of variation) on replicate assays of RiCof ranged from 9% to 34% or from 5% to 16% measuring the rate of aggregation. Similar values were found during a second international study undertaken for calibration of a new standard (T. W. BARROWCLIFf, personal communication). In our laboratory an interassay variability (measured as coefficient of variation) of 6% and 8.5% in low- and high-content vWf control plasma was obtained. Table 5 summarizes some technical suggestions for the performance of RiCof assay. Some authors prefer to quantitate the aggregation in RiCof assay by counting the platelets remaining after aggregation 5 or by measuring the time required to obtain a macroscopic clumping on a slide 2°. These methods have not gained wide popularity. 2.2 vWf'Ag assay Three principal methods are available for the quantitation of vWf:Ag, i.e., EIA (electroimmunoassay according to Laurell's technique), ELISA (enzymelinked immunosorbent assay) and IRMA (immunoradiometric assay). The sensitivity (lowest measurable level) is poor for EIA (5-10 U/dl), intermediate for ELISA (1-0.05 U/dl) using sandwich technique and polyclonal antibodies, and highest for IRMA (0.03-0.01 U/dl) using polyclonal antibodies 4,n,zg-31 The interassay precision has been estimated to be higher than 10% in Laurell's and IRMA assays and between 3.5% and 8% in ELISA4,n,~g-3L Even higher within-laboratory variability was reported in the above mentioned collaborative 149

T H E VONW'ILI~BRAND FACTOR

assay

advantages

disadvantages

EIA

inexpensive suitable for small laboratories easy to perform

long execution time automation not possible low sensitivity difficult standardization

IRMA

highest sensitivity

requires radioactive reagent handling requires technical expertness long execution time

ELISA

automation possible good sensitivity commercial kits available easy to perform short execution time

relatively high cost

T a b . 6 - A d v a n t a g e s a n d d i s a d v a n t a g e s o f t h e p r i n c i p a l a s s a y s f o r vWf:Ag.

studies for calibration of 1st and 2nd International Reference Calibration 2 (and T. W. BaRROWCLIFFE, personal communication). In our laboratory, using an ELISA method adapted for robotic instrumentation, an interassay variability o f 7% and 6% respectively has been obtained employing a control plasma with low and high vWf content 24. A strict correlation among the values estimated by these three methods has b e e n demonstrated both in normals and type I patients. Two different and intriguingly interrelated problems arise in type IIA patients. Firstly, ELISA and IRMA yield usually significantly lower vWf:Ag values, and secondly non-parallel dose-response curves have been observed for all the three methods n,29,3°. T h e first aspect could be anticipated considering that EIA is influenced both by molecular size and charge o f the investigated protein. The second aspect is not so clear, since it has b e e n found to occur also using Laurell's technique. Furthermore, within-family variability has been demonstrated for this p h e n o m e n o n and no correlation has been found between multimeric pattern and parallel or non-parallel dose-response curves in ELISA 3°. Thus, this aspect needs further investigation.

-

use highly binding, low optical absorbance microtitre plates use F(ab'), anti-vWf immunoperoxidase conjugate

-

record absorbance using double-wavelength reading

-

- optimize standard and sample dilution to work on the linear portion of the dose-response curve keep an accurate timing of dispensing reagent

-

test sample using at least two different dilutions

-

- replicate calibration curve -

use statistical approach in calculating results (parallel line bioassay)

- control plasma with low and high vWf content tested in each assay - calibrate your local standard against IRP automation advisable -

Tab. 7 - Technical suggestions for the performance

150

o f v W f : A g a s s a y b y ELISA.

F. RODEGHIERO, G. CASTAMAN

ELISA assay seems the most useful test for most clinical laboratories, as it appears from the analysis of advantages and disadvantages of the principal assays for vWf:Ag presented in tab. 6. Table 7 summarizes some technical suggestions for the performance of ELISA assay for vWf:Ag. CONCLUSIONS The available quantitative tests for the measurement of vWf in plasma have still a limited diagnostic sensitivity and the predictive value of an abnormal test is still limited. This makes the diagnosis of vWd troublesome and several tests are still required including RiCof, vWf:Ag, FVIII:C and bleeding time. Moreover, the investigations must often be repeated over time since a great variability is Observed and should be extended to several family members, especially in borderline cases whose prevalence is very high. At variance with clotting assay o f factor VIII, very few studies are available concerning the standardization of vWf assays and the limited data available offer a discouraging picture, especially concerning RiCof assay. Mulficenter collaborative studies seem to be urgently required to improve the performance o f these tests. SUMMARY Von Willebrand factor (vWf) is a multimeric and multivalent adhesive protein which is essential for platelet adhesion to subendothelium and for stabilization of factor VIII procoagulant activity in circulation. The quantitative measurement of vWf involves essentially two different approaches. The first is based on the interaction between vWf and Gp Ib of the platelet membrane in presence of ristocetin (ristocefin cofactor activity, RiCof) and depends not only on the amount of the factor but also on its ability to bring about this interaction, large multimers being more active. The second approach involves the immunological quantitafion of vWf (vWf:Ag) by its interaction with specific polyclonal or monoclonal antibodies as measured by several methods, i.e., electroimmunoassay, immunoradiometric assay and immunoenzymatic assay. Although in the majority of type II von Willebrand disease (vWd) with dysfunctional vWf there is a discrepancy between RiCof and vWf:Ag, it should be emphasized that RiCof activity does not entirely reflect the 'true' activity of vWf since it does not explore all the functions of this factor; furthermore, the relationship between degree of mulfimerization and RiCof level is not always tenable, as for example in vWd 'Vicenza'. For the diagnosis of congenital and acquired vWd RiCof assay together with family investigation is the eligible test, with an estimated ability to detect at least 50% of the carriers of the abnormal gene, including mildly affected patients; vWf:Ag appears less sensitive and, on the basis of studies carried out in our laboratory, a relative sensitivity of 64% is proposed. Both assays require the definition of separate normal ranges for children and adults and for 0 and non-0 blood group subjects; a nonparametric approach in a large sample of normal subjects is advisable. With RiCof assay performed by an aggregometric method using formalin-fixed platelets an interassay variability of 6% and 8.5% respectively for high- and low-control plasma was found in our laboratory. With vWf:Ag assayed by an ELISA method a variability of 7% for low- and 6% for high-control plasma was found. Thus, both methods appear sufficiently precise for clinical use. The use of an internal pool calibrated against an international standard allows to perform comparable interlaboratory measurements. To further improve standardization of these assays, collaborative studies seem urgently required. ACKNOWI Y.DGEMENTS This work was supported in part by grants from the Health Department of the Veneto Region (Progetti di Ricerca Sanitaria Finalizzata), and the Associazione V'wentina per l'Emofilia e le Coagulopatie. The technical assistance of Dr. A_ Tosetto is highly appreciated.

151

THE VON W/LLEBRAND FACTOR

REFERENCES 1. ABILDGAAm3C. F., SuzuKI Z., HARRISONJ., JEFFCOAT K., ZIMMERMANT. S.: Serial studies in von Willebrand's disease: variability versus 'variants' - Blood 56, 712, 1980. 2. BARROWCLIFFE T. W., TYDEMAN M. S., KIRKWOOD T. B. L., THOMAS D. P.: Standardization of factor VIII. III. Establishment of a stable reference plasma for factor VIII-related activities Thrombos. Haemostas. 50, 690, 1983. 3. BROWNJ. E., BOSAKJ. O.: An ELISA test for the binding of yon Willebrand antigen to collagen - Thrombos. Res. 43, 303, 1986. 4. CEJKAJ.: Enzyme immunoassay for factor VIii-related antigen - Clin. Chem. 28, 1356, 1982. 5. EVANS1L J., AtlSTEN D. E. G.: Assay of ristocetin co-factor using fixed platelets a n d a platelet counting technique - Brit.J. Haematol. 37, 289, 1977. 6. G I ~ J . C., ENImZs-BRooKs J., BAUER P.J., MARKS W.J., MONTGOMERY1L IL: T h e effect o f AB0 blood group on the diagnosis of yon Willebrand disease - Blood 69, 1691, 1987. 7. GIRMAJ. P., ME~R D., VERWEIJ C. L., PANNEXOEK H., SIX/VIAJ . j . : Structure-function relationship o f h u m a n yon WilIebrand factor - Blood 70, 605, 1987. 8. GI~_LNICKH. R., WILLIAMSS. B., McKEowN L. P., KIUZEK D. M., SHAFER B. C., RICK M. E.: Platelet yon Willebrand factor: comparison with plasma yon Willebrand factor - Thrombos. Res. 38, 623, 1985. 9. HANOIN 1L I., WAGNER D. D.: Molecular a n d cellular biology of yon Willebrand factor. In: COLLER B. S. (Ed.): Progress in Hemostasis a n d Thrombosis. W. B. Saunders Co., Philadelphia, 1989; vol. 9, p. 233. 10. HOYER L. W.: T h e factor VIII complex: structure a n d function - Blood 58, 1, 1981. 11. INGEaSlmVJ.: A sensitive ELISA for yon Willebrand factor (vWf:Ag) - Scan& J. clin. Lab. Invest. 47, t43, t987. 12. ITALIANWORKING GROUP: Spectrum of yon Willebrand's disease: a study of 100 cases - Brit. J. Haematol. 35, 101, 1977. 13. M A C F ~ E D. E., STIBBEJ., KIRBY E. P., ZUCm~RM. B , GP~mcr R. A_, McPHE~SON J.: A method for assaying yon Willebrand factor (ristocetin cofactor) - Thrombos. Diathes. haemorrh. (Stuttg.) 34, 306, 1975. 14. MANNUCCIP. M., LOMBARDI R., BADER R., VIANELLO L., FEDERICI A. B., SOLINASS., MAZZUCCONI M. G., MAmANI G.: Heterogeneity of type I yon Willebrand disease: evidence for a subgroup with an a b n o r m a l yon Willebrand factor - Blood 66, 796, 1985. 15. MANNUCCIP. M., LOMBAmgl1L, CASTAMANG., DENTJ. A., LATrUm3AA., ROOEGHIERO F., ZIMMERMAN T. S.: Von Willebrand disease 'Vicenza' with larger-than-normal (supranormal) yon Willebrand factor muldmers - Blood 71, 65, 1988. 16. MEYER D., JENKINS C. S. P., DRZYFUS M. D., FRESSlNAUD E., LARKIEU M.J.: Willebrand factor a n d ristocetin. II. Relationship between Willebrand factor, Willebrand antigen a n d factor V i i i activity - Brit_ J. Haematol. 28, 579, 1974. 17. MILLER C. H., GRAHAMJ. B., GOt.OlN L. R., ELSTON R. C.: Genetics of classic von Willebrand's disease. I. Phenotypic variation within families - Blood 54, 117, 1979. 18. MILLER C. H., GRAHAMJ. B., GOLDIN L. R., ElsroN R. C.: Genetics of classic von Willebrand's disease. II. Optimal assignment of the heterozygous genotype (diagnosis) by discriminant analysis - Blood 54, 137, 1979. t9. RANDJ. H., SUSSMANI. I., GORDON 1L E., CHU S. V., SOLOMON V.: Localization of factor-viiirelated antigen in h u m a n vascular subendothelium - Blood 55, 572, 1980. 20. REISNER H. M., KATZ H.J., GOLDIN L. R., BARROW E. S., GRAHAMJ. B.: Use of a simple visual assay of Willebrand factor for diagnosis a n d carrier identification - Brit. J. Haematol. 40, 339, 1978. 21. RODEGHIERO F., CASTAMANG.: Calibration o f lyophilized standards for ristocefin cofactor activity of von Willebrand factor (vW0 requires vWf-deficient plasma for dose-response curves Thrombos. Haemostas. 58, 978, 1987.

152

F. RODEGHIERO, G. CASTAMAN

22. RODEGHIERO F., CASTAMANG., DI BONA E., RUGGERI M., LOMBARDIR., MANNUCCI P. M.: Hyperresponsiveness to DDAVP for patients with type I v o n Willebrand's disease and normal intraptatelet yon Willebrand factor - Europ. J. Haematol. 40, 163, t988. 23. RODEGmERO K, CASTAMAN G., DINI E.: Epidemiological investigation of the prevalence of yon Willebrand's disease - Blood 69, 454, 1987. 24. RODEGHIERO F., CASTAMAZqG., TosErro A.: (Submitted for publication). 25. RUGGERI Z. M., DE MARCO L., GATrI L., BADER R., MONTGOMERY R. R_: Platelets have more t h a n o n e b i n d i n g site for yon Willebrand factor -J. clin. Invest. 72, 1, 1983. 26. RUGGERI Z. M., MANNUCCI P. M., BABER R., BARBUI T.: Factor VIII-related properties in platetets from patients with yon Willebrand's disease -J. Lab. clin. Med. 91, 132, 1978. 27. RUGGElU Z. M., MANNUCCI P. M., JEFFCOATE S. L., INGRAM G. I. C.: Immunoradiometric assay o f factor VIII-related antigen, with observations in 32 patients with yon Willebrand's disease BriL J. Haematol. 33, 221, 1975. 28. RUGGERI Z. M., ZlMMERM~ T. S.: Von Willebrand factor and yon Willebrand disease - Blood 70, 895, 1987. 29. SHOA'I I., LAVERGNEJ. M., A ~ L L O U N., OBERT B., ALA F., MEYER D.: Heterogeneity o f von Willebrand's disease: study of 40 Iranian cases - Brit. J. Haematol. 37, 67, 1977. 30. SHORT P. E., WILLIAMS C. E., ENAYATM. S., PICKEN A. M., HILL F. G. H.: Lack of correlation between factor VIii-related antigen multimeric analysis pattern a n d parallel or non-parallel dose response curves in an ELISA for factor VIii-related antigen assay -J. din. Pathol. 37, 194, 1984. 31. SHORT P. E., WILLIAMS C. E., PICKENA. M., HILL F. G. H.: Factor VIii-related antigen: a n improved enzyme immunoassay - Med. Lab. Sci. 39, 351, 1982. 32. SUSSMANI. I., RANDJ. H.: Subendothelial deposition of yon Willebrand's factor requires the presence o f endothelial ceils -J. Lab. clin. Med. 100, 526, 1982. 33. TscnoPP T. B., WEISS H.J., BAUMGARTNERH. tL: Decreased adhesion of platelets to subendothelium in yon Willebrand's disease -J. Lab. clin. Med. 83, 296, 1974. 34. TURITrO V. T., WEISS H.J., BAUMGARTNERH. R.: Ptatelet interaction with rabbit subendothelium in yon Willebrand's disease: altered t h r o m b u s formation distinct from defective platelet adhesion -J. clin. Invest. 74, 1730, 1984. 35. WEISS H.J., HOVER L. W., PICKLES F. R., WARMAA., ROGERSJ.: Quantitative assay of a plasma factor deficient in von Willebrand disease that is necessary for platelet aggregation -J. clin. I n v e s t 52, 2708, 1973. 36. WEISS H.J., SUSSMAN I- I., HOYER L. W.: Stabilization of factor VIII in plasma by the yon Willebrand f a c t o r - J , clin. Invest. 60, 390, 1977.

Requests for reprints should be addressed to: FRANCESCO RODEGHIERO

Divisione di Ematologia Centro Regionale per lo Studio delle Malattie Emorragiche e Trombotiche Ospedale San Bortolo 36100 Vicenza - Italia

153