Search

Platelet-associated IgG (PAIgG) has been reported to be elevated in nonthrombocytopenic patients who have a normal platelet lifespan. This has been interpreted as indicating that PAIgG is a nonspecific finding in these patients and not a determinant of platelet survival. It is important to recognize that the reticuloendothelial (RE) system plays an important role in the clearance of antibody-sensitized cells. In this study, we related the level of PAIgG and the platelet lifespan to the RE function in patients with: (A) idiopathic thrombocytopenic purpura (ITP), and (B) five patients with elevated levels of PAIgG yet normal or near-normal platelet counts. RE function was assessed by measuring the clearance of autologous chromium-labeled red cells sensitized with a precise amount of alloantibody (2,000–3,600 molecules of IgG/cell). Eight patients with immune thrombocytopenia had significantly shortened platelet survivals (less than 2–113 hr). In contrast, the five patients with elevated PAIgG, yet normal or near- normal platelet counts, all had normal autologous platelet survivals (186–222 hr). These patients also had significantly impaired clearance of IgG-sensitized red cells, with an average of 85% of the infused red cells remaining in the circulation at 60 min (normal 42% +/- 14%, n = 10). In this study, every patient with elevated PAIgG and normal RE function had a shortened platelet lifespan. Those patients with elevated PAIgG and impaired RE function did not invariably have a shortened platelet lifespan. The observation that the PAIgG is elevated in some patients whose platelet survival is normal does not indicate that PAIgG is not biologically relevant. It indicates that these patients may have RE blockade and do not clear IgG-sensitized cells.

Aspirin inhibits platelet function by acetylating platelet cyclooxygenase. Recent clinical trials indicate that aspirin is a promising antithrombotic agent against both venous and arterial thrombosis, but somewhat surprisingly this protective effect appears to be limited to males. To examine the potential sex-related differences in response to aspirin, we developed an animal model for quantitating fibrin accretion into an injury-induced thrombus and used it to study the effects of aspirin on thrombus size in male and female rabbits. Platelet prostaglandin synthesis was estimated by assay of platelet malondialdehyde and was significantly decreased in both male and female rabbits following treatment with 10 mg/kg aspirin (p less than 0.001). This inhibitory effect was not different for platelets from male and female rabbits. Thrombus size was significantly decreased in aspirin- treated male rabbits when compared to controls (p less than 0.05), but this aspirin effect was not apparent in female rabbits or rabbits of either sex treated with 10 mg/kg sodium salicylate. These findings support the results of clinical trials that were obtained by retrospective subgroup analysis. The reason for the sex difference is not known, but the findings raise an important issue in relationship to this mechanism of the antithrombotic effect of aspirin.

Thrombocytopenia is a common serious adverse effect of drug treatment. A variety of in vitro diagnostic techniques to confirm the diagnosis are available, but the majority lack sufficient sensitivity to detect all cases of drug-induced thrombocytopenia. We studied 19 patients with suspected drug-induced thrombocytopenia and demonstrated that platelet- associated IgG (PAIgG) was elevated in all at the time of thrombocytopenia, and PAIgG returned to normal levels as the thrombocytopenia resolved. In the majority of patients, the platelet count rapidly returned to normal after the drug was discontinued; however, in six patients, the thrombocytopenia persisted well beyond the period of time that the offending drug would be expected to be cleared from the blood. In 13 patients, serum obtained after recovery was used to identify the drug responsible for the thrombocytopenia in an in vitro assay. In all cases, the addition of the drug historically associated with the thrombocytopenic episode was associated with an increased binding of IgG to control platelets. For uncertain reasons, the concentration of drug required to increase the in vitro binding of IgG to test platelets was often more than the concentration usually achieved in vivo. Wider application of these techniques may provide better understanding of the clinical characteristics and mechanisms responsible for drug-induce thrombocytopenia.

The ability of platelets to adsorb vinblastine has been used to treat patients with immune thrombocytopenia. It is hypothesized that the drug- platelet complex is coated with antibody, taken up by macrophages which are then destroyed by the drug. We gave 16 courses of vinblastine- platelets to six patients with immune thrombocytopenia. Only one patient responded, and therefore we examined possible reasons for the lack of benefit. Using 3H-vinblastine, the kinetics of vinblastine binding to platelets was studied in vitro. The binding of vinblastine to both human and rabbit platelets was identical with maximal binding occurring within 10 min at 600 microgram/ml vinblastine. Similarly, the plasma half-life of vinblastine in rabbits was close to that reported for man, and therefore, in vivo binding of vinblastine to platelets in rabbits was considered a suitable model for man. Homologous donor rabbit platelets were labeled with 51Cr alone, 51Cr plus vinblastine, or 3H-vinblastine and infused into recipient rabbits. Vinblastine had no effect on 51Cr survival, but all measureable vinblastine had left the platelets within 2 hr of the infusion. These observations suggest that delivery of the vinblastine to the macrophages depends on the platelets being phagtocytized before the drug leaves the platelets. This would be likely to occur only in those patients with severe immune thrombocytopenia. Further investigations into this treatment should be directed at methods to maintain the drug within the platelet.

The mechanism of isolated thrombocytopenia in septicemia is unknown, but compensated disseminated intravascular coagulation (DIC) has been suggested as a possible cause. To investigate this possibility, platelet counts and sensitive assays for in vivo thrombin and plasmin generation, including fibrinogen gel chromatography and fibrinopeptide A (FPA) assays, were obtained on 31 septicemic patients. Fifteen of 17 patients with gram-negative septicemia and 8 of 14 patients with gram- positive septicemia had thrombocytopenia. Platelet survival studied demonstrated a decreased platelet survival. In 11 of 12 patients with severe thrombocytopenia (platelet count less than 50,000mul), there was laboratory evidence of intravascular coagulation. In contrast, there was little evidence of intravascular coagulation in 8 of 11 patients with moderate thrombocytopenia (platelet counts 50,000 to less than 150,000/mul) or in 7 of 8 patients with normal platelet counts. This report indicates that while DIC accompanies thrombocytopenia in many patients with severe thrombocytopenia, there is frequently little evidence for intravascular coagulation in patients with moderate thrombocytopenia. It is apparent that factors other than intravascular thrombin must play a role in producing the thrombocytopenia of septicemia.

Many patients with thrombotic thrombocytopenic purpura (TTP) have a platelet aggregating factor in their serum that may be pathologically linked with the disease process. To help characterize the type of platelet aggregation and platelet release induced by the sera from seven TTP patients, we measured the ability of a variety of inhibitors of platelet function as well as the ability of monoclonal antibodies (MoAbs) against platelet glycoproteins to inhibit TTP sera-induced platelet aggregation and release. These results were compared with the ability of the same inhibitors to block platelet aggregation induced by ristocetin, collagen, ADP, thrombin, and IgG-immune complexes. Monoclonal antibody directed against platelet glycoprotein Ib totally inhibited ristocetin-induced aggregation and release but had no effect on aggregation and release induced by the TTP sera or by any of the other platelet agonists. However, the MoAb against glycoproteins IIb/IIIa inhibited aggregation and release caused by TTP sera as well as by collagen, thrombin, and ADP but had no effect on aggregation and release induced by ristocetin. The aggregating activity could be abolished by heparin but not by the serine protease inhibitor PMSF (1 mmol/L). And although monomeric human IgG and purified Fc fragments of IgG inhibited IgG-immune complex-induced aggregation and release, they had no effect on TTP sera-induced aggregation and release nor on aggregation and release induced by any of the other agonists. Consistent with these in vitro studies showing no effect of IgG were the in vivo observations that intravenous (IV) IgG was without effect when administered to three patients with TTP. This study indicates that although a von Willebrand factor (vWF)-rich preparation of cryoprecipitate enhances the in vitro platelet aggregation and release caused by sera from the seven TTP patients we studied, the pathway of aggregation and release is not via platelet glycoprotein Ib. Also the aggregating factor of TTP sera is not neutralized in vitro or in vivo by IgG.

We have evaluated the fibrinogen/fibrin fragment E antigen assay as a diagnostic test in patients with clinically suspected venous thrombosis by comparing the results of this assay with venography in 272 patients. The result of the fragment E antigen assay was elevated in 79 of 80 patients with positive venograms for recent venous thrombosis (sensitivity 99%) and within the normal range in 161 of 192 patients with normal venograms (specificity 84%). The fragment E assay was also evaluated in 130 medical and surgical controls without evidence of venous thrombosis by leg scanning and the test was found to be relatively nonspecific. However, in the patient group under study, a correct clinical diagnosis of no thrombosis, based on a normal fragment E result, was made in 161 of 162 cases (negative predictive value of 99%). Therefore, a normal test result effectively excludes a diagnosis of venous thrombosis in clinically symptomatic patients. The assay, as currently performed, is technically demanding and takes 24 hr to complete. Therefore, it will have to be simplified before it can be applied to clinical practice.

Platelet-associated IgG (PAIgG) is elevated in idiopathic thrombocytopenic purpura (ITP), but it also is elevated in other thrombocytopenic disorders traditionally considered to be nonimmune. Consequently it is possible that elevated PAIgG is a nonspecific finding secondary to thrombocytopenia. To study this issue we developed a rabbit model of immune and nonimmune mediated thrombocytopenia. The mechanism of the thrombocytopenia was validated by platelet survival studies. Immune thrombocytopenia was produced by injection of antirabbit platelet serum that was raised in guinea pigs. Nonimmune aregenerative thrombocytopenia was produced by irradiation of the animals; nonimmune consumptive thrombocytopenia was produced by injection of adenosine diphosphate (ADP). PAIgG was measured in a direct binding assay using 125I-labeled staphylococcal protein A (SpA). Washed platelets from normal, nonthrombocytopenic rabbits bound an average of 81 molecules of SpA per platelet (81 +/- 168, mean +/- 2 SD, n = 39). Infusion of the antiplatelet antiserum produced thrombocytopenia with a rise in PAIgG that was closely correlated with the level of PAIgG (r = 0.86, n = 12). The thrombocytopenia was consumptive, as shown by a very short platelet life span using 111In- labeled platelets. In contrast, both nonimmune thrombocytopenic states resulted in an equal or greater drop in the platelet count but no change in the level of PAIgG. The animals with aregenerative thrombocytopenia had normal or only moderately reduced platelet life spans; however, in every animal the level of PAIgG was not different from the nonthrombocytopenic controls, irrespective of the platelet count. Similarly, the level of PAIgG was unchanged in those rabbits with nonimmune consumptive thrombocytopenia following infusion of ADP (82 +/- 55 molecules of SpA per platelet, mean +/- SD, n = 6). These studies indicate that elevated PAIgG is a specific finding of immune thrombocytopenia and is not secondary to thrombocytopenia itself. Indirectly these results support our hypothesis that immune mechanisms contribute to more thrombocytopenic disorders than was once thought likely.

Evidence suggests that as platelets age in the circulation they become progressively smaller and less dense through the loss of protein. The smallest, least dense platelets have a significantly shortened survival, but the mechanism of clearance of these platelets is not known. To evaluate whether the binding of IgG could play a role in the clearance of senescent platelets, we measured platelet size, total protein, and platelet-associated IgG on subpopulations of platelets isolated from 6 healthy individuals using a discontinuous iso-osmotic arabinogalactan (stractan) gradient. There was a close correlation between density, size, and total protein content (r greater than 0.9) for all platelet fractions. There was also a relationship between the amount of platelet-associated IgG (PAIgG), total protein, and platelet size (r greater than 0.9) for the first 3 progressively less dense platelet factions. However, the fourth platelet fraction containing the smallest, least dense, and on current evidence, oldest platelets had very elevated amounts of IgG. This amount was approximately 10 times higher than the mean platelet IgG for the same individual and was similar to the amount of PAIgG found on platelets from patients with immune thrombocytopenia. A progressive increase in the ratio of PAIgG measured after platelet solubilization to PAIgG measured on intact platelets was also noted for the first three populations, indirectly suggesting that platelets clear IgG from their surface during aging. Increased binding of IgG to senescent platelets may mediate their destruction.

Plasma and serum from patients with thrombotic thrombocytopenic purpura (TTP) can cause activation and aggregation of normal human platelets in vitro. It is possible that this platelet-activating factor contributes to the disease. In this report we describe studies designed to identify the platelet-activating factor in TTP. Platelet activation by sera from 15 patients with TTP was inhibited by leupeptin, iodoacetamide, and antipain but not by phenylmethylsulphonylfluoride, epsilon-aminocaproic acid, soybean trypsin inhibitor, aprotinin, and D-phenylanyl-1-prolyl-1- arginine chloromethyl ketone. These studies suggested that the platelet- activating factor in TTP serum was a cysteine protease. We confirmed that a calcium-dependent cysteine protease (CDP) was present in the sera of each of the 15 patients when we used an assay based on the ability of CDP to proteolyse platelet membrane glycoprotein 1b (GP1b) and hence to abolish the ability of CDP-treated normal platelets to agglutinate in the presence of ristocetin and von Willebrand factor. This proteolytic activity was inhibited by EDTA, leupeptin, antipain, iodoacetamide, and by N-ethyl-maleamide (NEM) but not by the serine protease inhibitors. Activity was detected in 15 of 15 patients with TTP tested before therapy was begun. In contrast, no activity was detected in the serum of any of five of the TTP patients tested in remission or in any of the sera from 36 patients with thrombocytopenia and 423 nonthrombocytopenic controls. To look for in vivo CDP activity in patients with TTP, we studied platelets from two patients with acute TTP (drawn into acid-citrate-dextrose, NEM, and leupeptin). These platelets showed a loss of GP1b from the platelet surface. Both patients were also studied in remission: GP1b on the platelet surface had returned to normal. These studies provide evidence that CDP is present in the sera of patients with TTP, that it is specific to this disease, and that is is active in vivo as well as in vitro. We postulate that a disorder of CDP homeostasis plays a major role in the pathophysiology of TTP.

Thrombocytopenia is a common occurrence (20%) in sick neonates, but the causes have not been well studied. In this report we demonstrate that thrombocytopenia in the neonate is characterized by increased platelet destruction as shown by shortened homologous 111In-oxine-labeled platelet life spans. Thirty-one prospectively studied thrombocytopenic neonates were investigated by measuring the 111In-labeled platelet life span, platelet-associated IgG (PAIgG), and coagulation screening tests. In every infant, the thrombocytopenia was shown to have a destructive component since the mean platelet life span was significantly shortened to 65 +/- 6 (mean +/- SEM) hours with a range of one to 128 hours compared with adult values (212 +/- 8; range, 140 to 260; gamma function analysis). The platelet survival was directly related to the lowest platelet count and inversely related to both the highest mean platelet volume and duration of the thrombocytopenia. In 22 infants the percent recovery of the radiolabeled platelets was less than 50%, which suggested that increased sequestration also contributed to the thrombocytopenia. Infants with laboratory evidence of disseminated intravascular coagulation (n = 8) or immune platelet destruction evidenced by elevated levels of PAIgG (n = 13) had even shorter platelet survivals and a more severe thrombocytopenia compared with the ten infants in whom an underlying cause for the thrombocytopenia was not apparent. Full-body scintigraphic images obtained in 11 infants showed an increased uptake in the spleen and liver, with a spleen-to- liver ratio of 3:1. This study indicates that thrombocytopenia in sick neonates is primarily destructive, with a subgroup having evidence of increased platelet sequestration.

Defective Fc receptor-specific reticuloendothelial (RE) function has been reported in certain patients with a variety of immunologic and nonimmunologic diseases. The mechanism responsible for the impaired RE function is uncertain, but it could be caused by immune complexes that are present in many of these disorders. Alternatively, the impaired RE function could be a secondary effect of the high concentrations of monomeric IgG in the serum of these patients, since monomeric IgG can compete with complexed IgG for macrophage receptors in vitro. We studied the Fc-dependent RE function in 30 healthy control subjects and 27 patients using IgG-coated radiolabeled autologous red cells. There was a significant relationship between the concentration of IgG in the serum and the rate of clearance of antibody-sensitized cells (r = 0.51, P less than .01). Patients with hypergammaglobulinemia had the slowest Fc-dependent clearance, whereas those with hypogammaglobulinemia had the most rapid clearance. Immune complexes (Raji or polyethylene glycol) could not be shown to contribute to Fc-dependent RE clearance above the effect of the IgG in the serum. The unusually rapid clearance in a patient with hypogammaglobulinemia could be returned to normal by raising the concentration of IgG in the serum. This study supports the concept that serum (monomeric) IgG competes with immune complexed IgG for macrophage Fc receptors in vivo. The competition for Fc receptors determines the level of competence of Fc-dependent RE function. Based on the results of this study, one can predict that a number of disorders characterized by hypergammaglobulinemia also will have impaired Fc-dependent RE function.

Heparin-induced thrombocytopenia can be a serious and difficult-to- diagnose complication of heparin therapy. Serum from patients with heparin-induced thrombocytopenia can cause heparin-dependent platelet aggregation, but the low sensitivity and specificity of this test limit its clinical usefulness. In this report we describe an assay for heparin-induced thrombocytopenia that is both sensitive and specific. The improvement in the assay was accomplished by measuring platelet release instead of aggregation and by measuring platelet release at two heparin concentrations. The rationale for the use of two heparin concentrations was that sera from patients with heparin-induced thrombocytopenia caused release at therapeutic but not at high concentrations of heparin. Twenty-eight sera samples from patients suspected of having heparin-induced thrombocytopenia and 573 controls were coded and tested in the assay. The patients with possible heparin- induced thrombocytopenia were ranked according to the likelihood of having this disorder by using prospectively defined criteria. The test had a high specificity (99%); only one of 573 controls showed a positive result. The test was also very sensitive, and the likelihood of a positive test result was directly correlated with the clinical likelihood of the patient having heparin-induced thrombocytopenia. Six of six patients with definitive heparin-induced thrombocytopenia had positive test results, whereas zero of four patients in whom the diagnosis was unlikely had positive test results. The two-point test for heparin-induced thrombocytopenia represents a sensitive and specific test for this disorder. This test may be useful not only in confirming the diagnosis of this disorder but also may provide information about its pathogenesis.

The biologic relevance of the increased platelet-associated IgG (PAIgG) on platelets from patients with idiopathic thrombocytopenic purpura (ITP) is unclear. Platelets from ITP patients are often larger than normal, and it is possible that the increased IgG is not specific but passively related to platelet size. The measurement of platelet-bound albumin could provide information concerning the specificity of the platelet-bound IgG, since albumin, like IgG, is a plasma protein, but unlike IgG, is not an active participant in immunologic reactions. Albumin is also a normal constituent of platelet membrane, and increased platelet albumin could indicate an increased platelet mass. Platelet-bound albumin, IgG, and total platelet protein were measured on both intact and disrupted platelets from healthy individuals (n = 25) and patients with ITP (n = 21). Platelet IgG and albumin were measured in an immunoradiometric assay using intact antisera and F(ab')2 fragments prepared from the same antisera. There was no relationship between platelet-bound IgG or albumin, and platelet size measured by either platelet protein or platelet volume, (r less than 0.3 for all interactions). In contrast, there was a significant correlation between platelet-bound albumin and platelet-bound IgG (r = 0.7, n = 21, p less than 0.001). Those patients with elevated platelet PAIgG also had elevated platelet albumin, and this relationship was irrespective of the total platelet protein content or mean platelet volume. It is possible that the increased platelet-bound IgG in ITP reflects an increase in platelet surface area or contaminating platelet fragments that are not manifested as an increase in platelet volume or total platelet protein. Alternatively, a platelet membrane abnormality may occur in ITP that results in the uptake of significant amounts of plasma proteins. Either possibility implies that not all of the IgG on platelets from patients with ITP is pathologic IgG.

The measurement of platelet-associated IgG (PAIgG) is a potentially useful diagnostic test for idiopathic thrombocytopenia purpura (ITP). However, the restricted application of PAIgG measurements to thrombocytopenic populations primarily comprised of ITP patients will artificially enhance its diagnosis specificity. For this reason, we performed a prospective study in which the results of a sensitive technique for quantitating PAIgG were related to the cause of the thrombocytopenia. Over a 1-yr period, clinicians were invited to submit patient blood samples encompassing as wide a spectrum of thrombocytopenic disorders as possible for PAIgG measurements. The physician was then contacted and requested to indicate the likeliest cause for the thrombocytopenia. The PAIgG was elevated in only 24 of 254 samples obtained from nonthrombocytopenic patients. In contrast, 134 (79%) of the 169 thrombocytopenic patients had elevated PAIgG results, and the increased levels were apparent in all diagnostic categories. The sensitivity of the PAIgG test for clinically diagnosed idiopathic thrombocytopenic purpura was 91% and the specificity was 27%. The positive predictive value for a raised PAIgG as a diagnostic test for ITP in a thrombocytopenic patient was only 46%, while the negative predictive value was 82%. This study indicates that the presence of increased PAIgG provides little additional information in the diagnosis of ITP. This study also suggests that immune mechanisms may mediate many more thrombocytopenic disorders than have been previously thought likely.

We observed several patients with chronic idiopathic thrombocytopenic purpura (ITP) whose bleeding times were more prolonged than would have been expected from their platelet counts. To investigate this further, we performed in vivo and in vitro platelet function studies, assessed arachidonate metabolism, and measured platelet-associated IgG (PAIGG) in seven patients with chronic ITP. The bleeding times of three of the patients were prolonged for greater than 7 min, and all of these patients had impaired platelet aggregation and abnormal platelet arachidonic acid metabolism as reflected by increased production of the lipoxygenase product HETE and a concomitant decrease in cyclooxygenase products, TXB2 and HHT (p less than 0.001). The abnormalities noted were not due to concomitant drug ingestion, since they were present on repeated evaluation. There was no relationship between the platelet count and the bleeding time; however, there was a significant inverse correlation between the bleeding time and TXB2 production in all patients evaluated (r = 0.81; p less than 0.05). There was no relationship between the level of platelet-associated IgG and any parameter of platelet aggregation or arachidonate metabolism. The abnormalities noted should be looked for in the individual patient with chronic ITP, since the bleeding tendency is exacerbated by the superimposed impairment of platelet function even at platelet counts of greater than 50,000/cu mm, levels generally regarded as “safe”.

We have compared the hemorrhagic and antithrombotic effects of a low molecular weight (LMW) heparin fraction and standard heparin in rabbits. Similar LMW heparin fractions have antithrombotic effects when tested in animals, but their hemorrhagic effects relative to standard heparin have not been established. Standard porcine mucosal heparin (mol wt 15,000 daltons) was depolymerized by nitrous acid to a low molecular weight fraction (mol wt 4600 daltons). Using equal USP units, the standard and Dep LMW heparin were compared in vitro, ex vivo, and in vivo. In vitro, when diluted in rabbit plasma, the Dep LMW heparin at equivalent anti-Xa activity showed less prolongation of thrombin clotting times or activated partial thromboplastin times. Ex vivo, platelets from rabbits treated with the Dep LMW heparin showed less inhibition of collagen-induced aggregation. The relative hemorrhagic properties of the two heparins were compared in vivo in rabbits using a sensitive blood loss assay, and the antithrombotic properties were compared in a thrombin-induced venous stasis model. By using an optimal threshold heparin dose in each test system, it was possible to demonstrate that equal USP units of Dep LMW heparin caused less blood loss but showed greater antithrombotic activity than standard heparin.

A number of methods have been developed to measure platelet-associated IgG (PAIgG). The antiglobulin consumption assay directly quantitates IgG on the platelet and is sensitive and specific. A fluorescent anti- IgG assay has recently been described and has the advantage of simplicity. We compared the results of these two PAIgG assays in immune and nonimmune thrombocytopenia and nonthrombocytopenic controls. The antiglobulin consumption assay was negative in 61 of 62 and the fluorescent negative in 54 of 62 assays in nonthrombocytopenic controls, and they were negative in 11 of 13 and 8 of 13 assays, respectively, in nonimmune thrombocytopenic patients. The antiglobulin consumption assay was positive in 54 of 58 and the fluorescent positive in 24 of 58 assays of patients with immune thrombocytopenia (ITP and SLE). The overall sensitivity and specificity of the antiglobulin consumption assay was 94% and 95% and of the fluorescent assay was 44% and 82%.

This report describes studies into the pathophysiology of heparin- induced thrombocytopenia. The IgG fraction from each of nine patients with heparin-induced thrombocytopenia caused heparin-dependent platelet release of radiolabeled serotonin. Both the Fc and the Fab portions of the IgG molecule were required for the platelet reactivity. The platelet release reaction could be inhibited by the Fc portion of normal human or goat IgG, and patient F(ab')2, but not F(ab')2 from healthy controls. These results suggested that the Fab portion of IgG binds to heparin forming an immune complex and the immune complexes initiate the platelet release reaction by binding to the platelet Fc receptors. To directly challenge this hypothesis, we preincubated the serotonin-labeled platelets with the monoclonal antibody against the platelet Fc receptor (IV.3). This monoclonal antibody completely inhibited the release reaction caused by heparin and patient sera, as well as heat aggregated IgG, but did not block collagen or thrombin- induced platelet release. Heparin-dependent platelet release also could be inhibited in vitro by the addition of monocytes and neutrophils, but not by red cells, presumably because the Fc receptors on the phagocytic cells have a higher binding affinity for IgG complexes than do platelets. Platelets from patients with congenital deficiencies of specific glycoproteins Ib and IX (Bernard-Soulier syndrome) and IIb and IIIa (Glanzmann's thrombasthenia) displayed normal heparin-dependent release indicating that the release reaction did not require the participation of these glycoproteins. These studies indicate that heparin-induced thrombocytopenia is an IgG-heparin immune complex disorder involving both the Fab and Fc portion of the IgG molecule.