We studied the possibility that immune injury to endothelial cells may have a role in the development of thrombosis in some patients with heparin-associated thrombocytopenia. Serum samples from each of 27 patients who had this clinical diagnosis contained heparin-dependent platelet antibodies and deposited more than normal amounts of IgG, IgA, or IgM on endothelial cells, stimulating the production of tissue factor. Binding of immunoglobulins to endothelial cells was no longer detected when the patients were studied after heparin was withdrawn, but reappeared within several days upon reexposure to heparin in a patient who experienced a clinical recurrence. Binding of immunoglobulin to endothelial cells was partially reduced by the preadsorption of serum samples with heparin or heparan sulfate bound to Sepharose or by enzymatic cleavage of cell-bound heparan sulfate, and was augmented by the addition of heparan sulfate. Thus, serum from some patients with heparin-associated thrombocytopenia may contain antibodies that react with heparin bound to endothelial cells or with heparan sulfate synthesized by endothelial cells. Immune injury to both platelets and endothelial cells may play a part in the development of thrombosis in some patients after heparin therapy.

We studied the mechanism of platelet injury in 20 patients receiving heparin, three of whom became thrombocytopenic. Platelets from these three patients had increased levels of IgG and C3, which correlated with the presence of thrombocytopenia; their plasma caused the release of serotonin from normal platelets at concentrations of heparin within the usual therapeutic range. This reaction required IgG and an intact classic complement pathway. The 17 patients receiving heparin in whom thrombocytopenia did not develop had normal levels of platelet-associated IgG and C3. Plasma from 14 of these patients caused the release of serotonin, but only at heparin concentrations above the therapeutic range, in a reaction that also required IgG and complement. The addition of heparin to 15 normal plasma samples did not cause platelet injury in vitro. These studies indicate that heparin administration can be associated with complement-mediated platlet injury. The dose-dependent nature of this process may account for the occurrence of thrombocytopenia in some of these patients.

Danazol and vinblastine are effective in many patients with chronic immune thrombocytopenic purpura (ITP). To evaluate the mechanism of action of these drugs, we studied six consecutive patients with chronic ITP treated with danazol and one treated with vinblastine. All the patients responded clinically without a notable change in the level of platelet-associated IgG. Instead, the clinical response to therapy was associated with a decrease in the number of monocyte binding sites for monomeric IgG (Fc receptors). In one patient, clinical relapse was associated with a spontaneous 2.7-fold increase in the number of monocyte Fc (IgG) receptors, without a change in the level of platelet-associated immunoglobulin. A decrease in the number of monocyte Fc (IgG) receptors following vinblastine infusion was associated with a clinical remission. We conclude that the clinical course of ITP may be influenced by the expression of monocyte or macrophage Fc (IgG) receptors. Danazol and vinblastine may mediate their clinical effect, at least in part, by influencing the number of available Fc (IgG) receptors on phagocytic cells.

Neonatal thrombocytopenia is a potentially life-threatening complication of immune thrombocytopenic purpura (ITP). We followed 23 pregnant women who had either a history of ITP (11 women) or clinically active disease (12 women) to delineate the factors responsible for neonatal thrombocytopenia. No relation was observed between maternal and neonatal platelet counts (P greater than 0.5). Eleven women delivered thrombocytopenic children; antiplatelet antibodies were detectable in each mother, including five who were in clinical remission at delivery. The level of platelet-associated IgG in the mothers did not identify the neonates at risk for thrombocytopenia (P greater than 0.05). However, the level of maternal circulating antiplatelet antibody correlated with both the presence and the extent of neonatal thrombocytopenia (P less than 0.005). A discrepancy between maternal platelet count and maternal antibody level may be especially notable in mothers treated with steroids or splenectomy. Monitoring the level of circulating antiplatelet antibody may help in identifying and managing pregnant women with ITP at risk of delivering neonates with serious thrombocytopenia.

No abstract