Epileppsia, 33(1): 178-184, 1992 Raven Press, Ltd., New York 0 International League Against Epilepsy
Valproate Therapy Induces von Willebrand Disease Type I W. Kreuz, R. Linde, M. Funk, R. Meyer-Schrod, E. Foll, U . Nowak-Gottl, G. Jacobi, *Zs. Vigh, and *I. Scharrer Klinikum der Johann Wolfgang Goethe-Universitat, Zentrum der Kinderheilkunde, Abteilung fur Padiatrische Hamatologie und Onkologie, Abteilung fur Padiatrische Neurologie; and *Zentrum der Inneren Medizin, Frankfurt a m Main, Germany
Summary: To investigate the increased tendency of hemorrhage in patients receiving valproate (VPA) therapy, we studied coagulation parameters in 30 randomized children of a group of 83 children receiving antiepileptic drug (AED) therapy. Besides a reduction in fibrinogen concentration and platelet count, we observed a significant decrease in factor VIII-complex. A decrease in factor VI1I:C was noted in 33%, a decrease in von Willebrand factor (vWF:Ag) was noted in 83% and a decrease in ristocetin-cofxtor activity (vWF:Rcof) was noted in 66% of the children. We classified a von Willebrand syndrome type I in 67% of our patients receiving VPA therapy. Sixty-three percent of patients had a history of bleeding, and 23% had a prolonged bleeding time. We compared
our results with those of a control group and of a group of patients with congenital von Willebrand disease (vWD), from which patients with multimer types I1 and I11 were excluded. Because coagulation parameters in patients with congenital vWD are similar to those receiving AED therapy, we designated the increased tendency to hemorrhage as VPA-induced vWD. The decrease in coagulation parameters were not dependent on either VPA dose or period of administration. In patients receiving VPA therapy, this result must be considered, especially during surgical intervention and after traumatic events. Key Words: Anticonvulsants-Valproate, drug-induced abnormalities-Hemorrhage-von Willebrand syndromeChildren.
Since 1967 valproate (VPA) has been used extensively in therapy of generalized epilepsies and has proven as effective as carbamazepine (CBZ) or phenytoin (PHT) in the management of generalized tonic-clonic seizures (GTC). VPA may also be effective in therapy of infantile spasms (Anonymous, 1988). In 1989 VPA was administered to 45,000 patients in Germany and to 200,000 patients in the United States (Institute for Industrial Medicine, Frankfurt, 1989). With VPA, most patients remain symptom-free, but some have shown a severe hepatotoxicity, with a case-fatality rate of - 1 in 10,000 (Jacobi et al., 1980; Dreifuss et al., 1987, 1989). Hematologic reactions are another important side effect with several reports of hemorrhagic diathesis, thrombocytopenia, platelet dysfunction, and decreased fibrinogen on VPA administration (Sutor, 1976; Loiseau, 1981 ;Jeavons, 1982). We recently reported prelim-
inary results on induction of von Willebrand disease (vWD) by VPA (Kreuz et al., 1990). We now present complete data, including two case reports, on VPA-induction of vWD in children. PATIENTS AND METHODS
In our outpatient clinic, we frequently observed a history of hemorrhage in children treated with VPA. In a group of 83 children receiving VPAtherapy, we often noted hemorrhagic disorders combined with a decrease in factor VIII/vWFcomplex. Because these laboratory findings appeared to be similar to those observed in congenital vWD, we continued our observation in 30 randomized children. These children (aged 1-18 years, 16 male and 14 female), treated with VPA, were studied from March 1985 to January 1987. Total VPA was 300-2,500 mg/day, equivalent to 20-30 mg/kg body weight and, in rare patients, S120 mg/kg body weight. Therapy ranged from 6 months to 14 years. These children had not previously had a hemostatic defect and had no family history of bleeding disorders. The data in VPAtreated children were compared with those of a
Received December 1990; revision accepted January 1991. Address correspondenceand reprint requests to Dr. W. Kreuz at Klinikum der Johann Wolfgang Goethe-Universitat, Zentrum der Kinderheilkunde, Abteilung fur Padiatrische Hamatologie und Onkologie, Theodor-Stern-Kai 7, 6000 Frankfurt am Main 70, Germany.
178
VPA INDUCES vWD TYPE 1 TABLE 1. History of bleeding episodes Positive, n(%)
Group Children with congenital vWD type I (n = 43) Children receiving VPA (n = 30)
Negative, n (%)
TABLE 2. Multimer distribution in children with V P A therapy and in children with congenital (vWD)
Not known, n
23 (53)
17 (40)
3 (7)
19 (63)
1 1 (37)
-
179
Group Congenital vWD (n = 43) Children receiving VPA(n = 30)
VPA, valproate.
Normal, n (%I
Intermediate normal-type I, n (5%)
Type I, n (%I
6 (14)
6 (14)
31 (72)
4 (13)
6 (20)
20 (67)
VPA, valproate; vWD, von Willebrand disease
group of 43 children (30 male and 13 female aged 7 months to 14 years) with congenital vWD type I, excluding multimer types I1 and 111. Both groups were also compared with a control group (40 children, 18 male and 22 female, aged 6 months to 18 years) consisting of children without a bleeding disorder. Bleeding time was measured according to the Simplate I method (norm 2 rnin 30 s to 9 rnin 30 s). Factor VII1:C activity (VIII:C, norm 70-100%) was determined in a one-step chromogenic coagulation assay (Baxter Health Care Corp., Glendale, CA, U.S.A.). Von Willebrand factor (vWF:Ag, norm 70-1 50%) was determined by a quantitative electroimmunoassay (EIA) and ristocetin cofactor activity (vWF:Rcof, norm >70%) assessed using formalin-fixed platelets (Zuzel et al., 1978). The multimer distribution of vWF:Ag was assayed with sodium dodecyl sulfate (SDS)-agarose electrophoresis as described by Ruggeri and Zimmerman (1981). For data analysis as percentile, we used the last blood test in each patient. Therefore, in some of our congenital vWD patients values of factor VIII/ vWF-complex were normal as a result of having blood drawn while under stress (Ingram, 1961), but these patients had all shown prior decreased values. Statistical comparisons were made by the rank-sum test for unpaired data (Mann-Whitney U test).
tory of hemorrhagic diathesis (Table 1). Mostly, we observed mild bleeding episodes as skin hemorrhages, epistaxis and gum hemorrhages. The bleeding time in the group of children treated with VPA was prolonged to nearly the same extent as that in children with congenital vWD. The bleeding time in children receiving VPA ranged between 3 and 17 rnin (median 7 rnin 45 s). In this group, 23 children (77%) had a normal bleeding time and 7 children (23%) had a prolonged bleeding time. Children with congenital vWD had bleeding times ranging from 4 rnin 30 s to 21 rnin (median 9 rnin 30 s). Twenty-two children (51%) in this group had a normal bleeding time and 21 children (49%) had an elongated bleeding time (Fig. 1). The values of factor VII1:C were compared in the three groups (Fig. 2). The concentration of factor VII1:C in VPA-treated children ranged from 31 to 108% (median 80%). Ten children (33%) had reduced levels of factor VII1:C (
Valproate Therapy Induces von Willebrand Disease Type I W. Kreuz, R. Linde, M. Funk, R. Meyer-Schrod, E. Foll, U . Nowak-Gottl, G. Jacobi, *Zs. Vigh, and *I. Scharrer Klinikum der Johann Wolfgang Goethe-Universitat, Zentrum der Kinderheilkunde, Abteilung fur Padiatrische Hamatologie und Onkologie, Abteilung fur Padiatrische Neurologie; and *Zentrum der Inneren Medizin, Frankfurt a m Main, Germany
Summary: To investigate the increased tendency of hemorrhage in patients receiving valproate (VPA) therapy, we studied coagulation parameters in 30 randomized children of a group of 83 children receiving antiepileptic drug (AED) therapy. Besides a reduction in fibrinogen concentration and platelet count, we observed a significant decrease in factor VIII-complex. A decrease in factor VI1I:C was noted in 33%, a decrease in von Willebrand factor (vWF:Ag) was noted in 83% and a decrease in ristocetin-cofxtor activity (vWF:Rcof) was noted in 66% of the children. We classified a von Willebrand syndrome type I in 67% of our patients receiving VPA therapy. Sixty-three percent of patients had a history of bleeding, and 23% had a prolonged bleeding time. We compared
our results with those of a control group and of a group of patients with congenital von Willebrand disease (vWD), from which patients with multimer types I1 and I11 were excluded. Because coagulation parameters in patients with congenital vWD are similar to those receiving AED therapy, we designated the increased tendency to hemorrhage as VPA-induced vWD. The decrease in coagulation parameters were not dependent on either VPA dose or period of administration. In patients receiving VPA therapy, this result must be considered, especially during surgical intervention and after traumatic events. Key Words: Anticonvulsants-Valproate, drug-induced abnormalities-Hemorrhage-von Willebrand syndromeChildren.
Since 1967 valproate (VPA) has been used extensively in therapy of generalized epilepsies and has proven as effective as carbamazepine (CBZ) or phenytoin (PHT) in the management of generalized tonic-clonic seizures (GTC). VPA may also be effective in therapy of infantile spasms (Anonymous, 1988). In 1989 VPA was administered to 45,000 patients in Germany and to 200,000 patients in the United States (Institute for Industrial Medicine, Frankfurt, 1989). With VPA, most patients remain symptom-free, but some have shown a severe hepatotoxicity, with a case-fatality rate of - 1 in 10,000 (Jacobi et al., 1980; Dreifuss et al., 1987, 1989). Hematologic reactions are another important side effect with several reports of hemorrhagic diathesis, thrombocytopenia, platelet dysfunction, and decreased fibrinogen on VPA administration (Sutor, 1976; Loiseau, 1981 ;Jeavons, 1982). We recently reported prelim-
inary results on induction of von Willebrand disease (vWD) by VPA (Kreuz et al., 1990). We now present complete data, including two case reports, on VPA-induction of vWD in children. PATIENTS AND METHODS
In our outpatient clinic, we frequently observed a history of hemorrhage in children treated with VPA. In a group of 83 children receiving VPAtherapy, we often noted hemorrhagic disorders combined with a decrease in factor VIII/vWFcomplex. Because these laboratory findings appeared to be similar to those observed in congenital vWD, we continued our observation in 30 randomized children. These children (aged 1-18 years, 16 male and 14 female), treated with VPA, were studied from March 1985 to January 1987. Total VPA was 300-2,500 mg/day, equivalent to 20-30 mg/kg body weight and, in rare patients, S120 mg/kg body weight. Therapy ranged from 6 months to 14 years. These children had not previously had a hemostatic defect and had no family history of bleeding disorders. The data in VPAtreated children were compared with those of a
Received December 1990; revision accepted January 1991. Address correspondenceand reprint requests to Dr. W. Kreuz at Klinikum der Johann Wolfgang Goethe-Universitat, Zentrum der Kinderheilkunde, Abteilung fur Padiatrische Hamatologie und Onkologie, Theodor-Stern-Kai 7, 6000 Frankfurt am Main 70, Germany.
178
VPA INDUCES vWD TYPE 1 TABLE 1. History of bleeding episodes Positive, n(%)
Group Children with congenital vWD type I (n = 43) Children receiving VPA (n = 30)
Negative, n (%)
TABLE 2. Multimer distribution in children with V P A therapy and in children with congenital (vWD)
Not known, n
23 (53)
17 (40)
3 (7)
19 (63)
1 1 (37)
-
179
Group Congenital vWD (n = 43) Children receiving VPA(n = 30)
VPA, valproate.
Normal, n (%I
Intermediate normal-type I, n (5%)
Type I, n (%I
6 (14)
6 (14)
31 (72)
4 (13)
6 (20)
20 (67)
VPA, valproate; vWD, von Willebrand disease
group of 43 children (30 male and 13 female aged 7 months to 14 years) with congenital vWD type I, excluding multimer types I1 and 111. Both groups were also compared with a control group (40 children, 18 male and 22 female, aged 6 months to 18 years) consisting of children without a bleeding disorder. Bleeding time was measured according to the Simplate I method (norm 2 rnin 30 s to 9 rnin 30 s). Factor VII1:C activity (VIII:C, norm 70-100%) was determined in a one-step chromogenic coagulation assay (Baxter Health Care Corp., Glendale, CA, U.S.A.). Von Willebrand factor (vWF:Ag, norm 70-1 50%) was determined by a quantitative electroimmunoassay (EIA) and ristocetin cofactor activity (vWF:Rcof, norm >70%) assessed using formalin-fixed platelets (Zuzel et al., 1978). The multimer distribution of vWF:Ag was assayed with sodium dodecyl sulfate (SDS)-agarose electrophoresis as described by Ruggeri and Zimmerman (1981). For data analysis as percentile, we used the last blood test in each patient. Therefore, in some of our congenital vWD patients values of factor VIII/ vWF-complex were normal as a result of having blood drawn while under stress (Ingram, 1961), but these patients had all shown prior decreased values. Statistical comparisons were made by the rank-sum test for unpaired data (Mann-Whitney U test).
tory of hemorrhagic diathesis (Table 1). Mostly, we observed mild bleeding episodes as skin hemorrhages, epistaxis and gum hemorrhages. The bleeding time in the group of children treated with VPA was prolonged to nearly the same extent as that in children with congenital vWD. The bleeding time in children receiving VPA ranged between 3 and 17 rnin (median 7 rnin 45 s). In this group, 23 children (77%) had a normal bleeding time and 7 children (23%) had a prolonged bleeding time. Children with congenital vWD had bleeding times ranging from 4 rnin 30 s to 21 rnin (median 9 rnin 30 s). Twenty-two children (51%) in this group had a normal bleeding time and 21 children (49%) had an elongated bleeding time (Fig. 1). The values of factor VII1:C were compared in the three groups (Fig. 2). The concentration of factor VII1:C in VPA-treated children ranged from 31 to 108% (median 80%). Ten children (33%) had reduced levels of factor VII1:C (
