Intravenous immunoglobulin as an adjunct therapy in persisting heparin-induced thrombocytopenia.

Heparin induced thrombocytopenia (HIT) is a serious adverse drug reaction caused by transient antibodies against platelet factor 4 (PF4)/heparin complexes, resulting in platelet activation and potentially fatal arterial and/or venous thrombosis. Most cases of HIT respond to cessation of heparin and administration of an alternative non-heparin anticoagulant, but there are cases of persisting HIT, defined as thrombocytopenia due to platelet activation/consumption for greater than seven days despite standard therapy. These patients remain at high risk for thrombotic events, which may result in limb-loss and mortality. Intravenous immunoglobulin (IVIg) has been proposed as an adjunct therapy for these refractory cases based on its ability to saturate FcγRIIa receptors on platelets, thus preventing HIT antibody binding and platelet activation. We describe 2 cases of persisting HIT (strongly positive antigen and functional assays, and persisting thrombocytopenia >7 days) with rapid clinical response to IVIg. We performed in-vitro experiments to support IVIg response. Healthy donor platelets (1 × 10e6) were treated with PF4 (3.75 µg/mL) for 20 min followed by 1-hour incubation with patients' sera. Platelet activation with and without addition of IVIg (levels equivalent to those reached in a patient after treatment with 2 gm/Kg) was evaluated in the PF4-dependent P-selectin expression assay (PEA). A significantly decreased platelet activation was demonstrated after the addition of IVIg to both patient samples, which correlated well with the rapid clinical response that each patient experienced. Thus, our study supports the use of IVIg as an adjunct therapy for persisting HIT.

Glacial seismology.

Seismic source and wave propagation studies contribute to understanding structure, transport, fracture mechanics, mass balance, and other processes within glaciers and surrounding environments. Glaciogenic seismic waves readily couple with the bulk Earth, and can be recorded by seismographs deployed at local to global ranges. Although the fracturing, ablating, melting, and/or highly irregular environment of active glaciers can be highly unstable and hazardous, informative seismic measurements can commonly be made at stable proximal ice or rock sites. Seismology also contributes more broadly to emerging studies of elastic and gravity wave coupling between the atmosphere, oceans, solid Earth, and cryosphere, and recent scientific and technical advances have produced glaciological/seismological collaborations across a broad range of scales and processes. This importantly includes improved insight into the responses of cryospheric systems to changing climate and other environmental conditions. Here, we review relevant fundamental physics and glaciology, and provide a broad review of the current state of glacial seismology and its rapidly evolving future directions.

Prostaglandin E 1 in preparation and storage of platelet concentrates.

The addition of prostaglandin E(1) to blood collection bags improved platelet harvesting; platelets were, esily suspended immediately after centrifugation. The treatment with prostaglandin E(1) did not affect the survival time of these platelets after infusion into a recipient.

Deletion of the platelet-specific alloantigen PlA1 from platelets in Glanzmann's thrombasthenia.

Expression of a Platelet-specific alloantigen (Pl(A1)) was studied in five unrelated patients with Glanzmann's thrombasthenia using immunologic techniques based on release of (51)Cr from tagged platelets by Pl(A1)-specific antibody. Less than 1% of the normal quantity of Pl(A1) could be detected on platelets of patients 1, 2, and 3; platelets from patients 4 and 5 contained 22 and 12% of normal levels, respectively. After treatment with bromelain, platelets from patients 4 and 5, but not those from patients 1, 2, and 3, released (51)Cr as well as normal Pl(A1)-positive platelets when exposed to anti-Pl(A1). Platelets from each of the five patients reacted normally with drug-dependent antibodies and with autoantibodies specific for platelets. Polyacrylamide gel electrophoresis of thrombasthenic platelets showed marked deficiencies of glycoproteins IIbalpha and III (P < 0.0005), confirming recent reports of others. Deficiency of the two proteins as determined by gel scanning was more pronounced in patients 1, 2, and 3 than in patients 4 and 5. Normal levels of glycoproteins IIbalpha and III were found in platelets from normal subjects negative for Pl(A1). These observations are consistent with the possibility that the Pl(A1) antigen is located on one or both of the glycoproteins lacking in Glanzmann's thrombasthenia, although other explanations are possible. They further suggest that patients with thrombasthenia may be heterogeneous in respect to the degree to which these glycoproteins are deleted. The Pl(A1) antigen can be measured with considerable precision and may provide a marker useful for the diagnosis and study of Glanzmann's disease.

A platelet and granulocyte membrane defect in paroxysmal nocturnal hemoglobinuria: usefulness for the detection of platelet antibodies.

The tendency of platelets and leukocytes to lyse after their interaction with antibody and complement was studied by measuring the release of (51)Cr from cells labeled with this isotope. Platelets from six patients with paroxysmal nocturnal hemoglobinuria (PNH) were 15-230 times more sensitive to antibodies and 10-32 times more sensitive to complement than normal platelets or platelets from patients with other types of thrombocytopenic or hemolytic disorders. Mixed white blood cell (WBC) preparations from patients with PNH were 3-20 times more sensitive to anti-WBC antibodies and 5-10 times more sensitive to C' than were WBC preparations from normal subjects, but PNH lymphocytes showed normal immunologic reactivity. PNH platelets, like PNH erythrocytes, lysed more readily than normal platelets in acidified serum and in media of reduced ionic strength, but these characteristics were not demonstrable with PNH WBC's under the conditions of study. In PNH, platelets appear to comprise a single population with respect to their sensitivity to immune lysis, yet their survival time as measured with (51)Cr falls within normal limits. PNH granulocytes likewise appear to consist of a single, uniformly sensitive population. It is concluded that, in PNH, platelets and granulocytes share the membrane defect characteristic of erythrocytes in this disorder. These observations support the concept that PNH arises as the result of a somatic mutation in a primitive cell capable of differentiating into erythroblast, myeloblast, and megakaryoblast lines. PNH platelets or enzymatically treated normal platelets permit the detection of some types of platelet antibodies in dilutions up to 2000-fold greater than is possible with currently available methods, a finding suggesting that the immune lysis technique will prove useful for the study of platelet immunology.

Drug-antibody-platelet interaction in quinine- and quinidine-induced thrombocytopenia.

Binding of quinine- and quinidine-dependent antibodies to platelets was studied using an electroimmunoassay to measure platelet-bound IgG. Antibodies from four patients with drug-induced thrombocytopenia differed significantly in their interaction with platelets: association constants for binding to platelets at high drug concentrations ranged from 0.29 to 2.6 x 10(7) M(-1), the maximum number of antibody molecules bound ranged from 36,000 to 161,000/platelet, the amount of drug necessary to achieve half-maximum binding of antibodies to platelets ranged from 2 to 60 muM, and only one of the antibodies cross-reacted with the stereoisomer of the drug to which the patient was sensitized. Binding of the antibodies to platelets was enhanced at the highest achievable molar ratio of drug:antibody, 10,000:1, rather than being inhibited, as would be expected in a conventional, hapten-dependent reaction. The drug-antibody-platelet reaction was unaffected by Factor VIII/von Willebrand protein, nonspecifically aggregated IgG, or heat-labile complement components. After pretreatment with tritiated quinine, platelets retained several hundred thousand molecules of drug each, but failed to bind detectable amounts of antibody. However, platelets treated simultaneously with quinine-dependent antibody and tritiated quinine retained significantly more drug after repeated washes than platelets treated with drug and normal serum. These findings support the proposition that in quinine- and quinidine-induced thrombocytopenia, drug and antibody combine first in the soluble phase to form a complex, which then binds with high affinity to a receptor on the platelet surface (innocent bystander reaction), and demonstrate that these antibodies are heterogeneous in respect to the amount of drug required to promote their binding to platelets, the number of platelet receptors they recognize, and their binding affinities.

Fab-mediated binding of drug-dependent antibodies to platelets in quinidine- and quinine-induced thrombocytopenia.

Platelets coated with quinine- or quinidine-induced antibodies form rosettes around protein A-Sepharose beads and normal platelets form rosettes about protein A-Sepharose beads coated with these antibodies. These reactions occurred only in the presence of sensitizing drug. Platelets also formed rosettes about protein A-Sepharose beads coated with an anti-PIA1 antibody, but drug was not required. Formation of rosettes between antibody-coated platelets and protein A-Sepharose was inhibited by F(ab')2 fragments of goat antibody specific for the Fc portion of human IgG, while rosette formation between antibody-coated protein A-Sepharose and platelets was inhibited by F(ab')2 fragments directed against the F(ab')2 portion of the IgG molecule. Since binding of IgG to protein A is known to occur via the Fc region, these findings suggest that binding of drug-induced antibodies to platelets occurs at the Fab domains of the IgG molecule.

Absence of the platelet receptor for drug-dependent antibodies in the Bernard-Soulier syndrome.

The platelet membrane receptor for quinidine- and quinine-dependent antibodies was studied in three patients with the Bernard-Soulier syndrome (BSS) and in normal subjects with immunologic techniques based on the release of 51Cr from labeled platelets. The receptor could not be detected on BSS platelets but was present on platelets from each of 180 normal subjects. BSS platelets reacted normally with other allo- and autoantibodies. In confirmation of previous reports, BSS platelets were found to be deficient in glycoproteins Ib and Is. However, after apparently total cleavage of these proteins from the membrane of normal platelets by controlled hydrolysis with trypsin or chymotrypsin, 80% of the drug-dependent antibody receptor activity was retained. These observations suggest the existence of an additional, hitherto unrecognized membrane defect in Bernard-Soulier platelets.

The human platelet alloantigens, PlA1 and PlA2, are associated with a leucine33/proline33 amino acid polymorphism in membrane glycoprotein IIIa, and are distinguishable by DNA typing.

The human platelet alloantigens, PlA1 and PlA2, comprise a diallelic antigen system located on a component of the platelet fibrinogen receptor, membrane glycoprotein (GP) IIIa. Of the known platelet alloantigens, PlA1, which is carried by 98% of the caucasian population, appears to be the alloantigen that most often provokes neonatal alloimmune thrombocytopenic purpura and posttransfusion purpura. The structural features of the GPIIIa molecule responsible for its antigenicity are as yet unknown. Using the polymerase chain reaction (PcR), we amplified the NH2-terminal region of platelet GPIIIa mRNA derived from PlA1 and PlA2 homozygous individuals. Nucleotide sequence analysis of selected amplified cDNA products revealed a C in equilibrium T polymorphism at base 196 that created a unique Nci I restriction enzyme cleavage site in the PlA2, but not the PlA1 form of GPIIIa cDNA. Subsequent restriction enzyme analysis of cDNAs generated by PcR from 10 PlA1/A1, 5 PlA2/A2, and 3 PlA1/A2 individuals showed that Nci I digestion permitted clear discrimination between the PlA1 and PlA2 alleles of GPIIIa. All PlA2/A2 individuals studied contain a C at base 196, whereas PlA1 homozygotes have a T at this position. This single base change results in a leucine/proline polymorphism at amino acid 33 from the NH2-terminus, and is likely to impart significant differences in the secondary structures of these two allelic forms of the GPIIIa molecule. The ability to perform DNA-typing analysis for PlA phenotype may have a number of useful clinical applications, including fetal testing and determination of the phenotype of severely thrombocytopenic individuals.

Antibodies from patients with heparin-induced thrombocytopenia/thrombosis are specific for platelet factor 4 complexed with heparin or bound to endothelial cells.

Heparin-induced thrombocytopenia/thrombosis (HITP) is thought to be mediated by immunoglobulins that activate platelets in the presence of pharmacologic concentrations of heparin, but the molecular basis for this relatively common and often serious complication of heparin therapy has not been established. We found that plasma from each of 12 patients with HITP contained high titer (> or = 1:200) antibodies that reacted with immobilized complexes of heparin and platelet factor 4 (PF4), a heparin-binding protein contained in platelet alpha-granules. Recombinant human PF4 behaved similarly to PF4 isolated from platelets in this assay system. Complexes formed at an apparent heparin/PF4 molecular ratio of approximately 1:2 (fresh heparin) and approximately 1:12 (outdated heparin) were most effective in binding antibody. Immune complexes consisting of PF4, heparin, and antibody reacted with resting platelets; this interaction was inhibited by a monoclonal antibody specific for the Fc gamma RII receptor and by excess heparin. Human umbilical vein endothelial cells, known to express heparin-like glycosaminoglycan molecules on their surface, were recognized by antibody in the presence of PF4 alone; this reaction was inhibited by excess heparin, but not by anti-Fc gamma RII. Antibodies reactive with heparin/PF4 were not found in normal plasma, but IgG and IgM antibodies were detected at dilutions of 1:10 (IgG) and 1:50 (IgM) in 3 of 50 patients (6%) with other types of immune thrombocytopenia. These findings indicate that antibodies associated with HITP react with PF4 complexed with heparin in solution or with glycosaminoglycan molecules on the surface of endothelial cells and provide the basis for a new hypothesis to explain the development of thrombocytopenia with thrombosis or disseminated intravascular coagulation in patients sensitive to heparin.

On the association of the platelet-specific alloantigen, Pena, with glycoprotein IIIa. Evidence for heterogeneity of glycoprotein IIIa.

Neonatal alloimmune thrombocytopenic purpura associated with a new platelet-specific alloantigen Pena has been reported. We now provide direct evidence that the Pena determinant is associated with glycoprotein (GP) IIIa, but that it is distinct from epitopes that define the PlA system. By ELISA wherein monoclonal antibodies specific for GPIIb (Tab) and specific for GPIIIa (AP3) were used to capture and hold antigens from a platelet lysate prepared under conditions that generate free GPIIb and GPIIIa, anti-Pena reacted with GPIIIa held by AP3 but not with GPIIb held by Tab. In an alternative ELISA where purified GPIIIa from both PlA1-positive and PlA1-negative platelets were used individually as antigen, anti-Pena reacted with both allelic forms of GPIIIa. By radioimmuno-precipitation, anti-Pena precipitated a single surface-labeled membrane protein with electrophoretic characteristics in sodium dodecyl sulfate-polyacrylamide gels, under nonreduced or reduced conditions, identical to those of GPIIIa. By fluorocytometry, platelets from several donors, regardless of PlA phenotype, bound an amount of anti-Pena roughly equivalent to one-half that amount of anti-PlA1 bound by PlA1 homozygous (A1/A1) platelets and roughly equal to that amount of anti-PlA1 bound by PlA1 heterozygous (A1/A2) platelets. Using platelets from donors typed homozygous for PlA1 and Pena in a quantitative indirect binding assay, 14-24,000 molecules of anti-Pena and 41-51,000 molecules of anti-PlA1 were bound per platelet at saturation. Anti-Pena completely inhibited ADP-induced aggregation of Pena-positive platelets, regardless of PlA phenotype. These results indicate that the Pena determinant is associated with GPIIIa but distinct from PlA.

An amino acid polymorphism within the RGD binding domain of platelet membrane glycoprotein IIIa is responsible for the formation of the Pena/Penb alloantigen system.

The human Pena/Penb alloantigen system represents a naturally occurring polymorphism of human platelet membrane glycoprotein (GP) IIIa, and has previously been implicated in the onset of two important clinical syndromes, neonatal alloimmune thrombocytopenic purpura and posttransfusion purpura. To investigate the molecular basis of the polymorphism underlying the Pen alloantigen system, we used the polymerase chain reaction to amplify platelet-derived GPIIIa mRNA transcripts. DNA sequence analysis of amplified GPIIIa cDNAs from nucleotides 161 to 1341 (encompassing amino acid residues 22-414) revealed a G526<==>A526 polymorphism that segregated precisely with Pen phenotype in twelve other individuals examined. This nucleotide substitution results in an Arg (CGA) to Gln (CAA) polymorphism at amino acid 143 of GPIIIa. Interestingly, this polymorphic residue is located within the putative RGD binding site (residues 109-171) of GPIIIa. Platelet aggregation patterns of a Penb/b individual, however, were nearly normal in response to all physiological agonists tested, indicating that this polymorphism does not grossly affect integrin function. Short synthetic peptides encompassing residue 143 were unable to mimic either the Pena or Penb antigenic determinants, suggesting that the Pen epitopes are dependent upon proper folding of the polypeptide chain. Finally, we constructed allele-specific recombinant forms of GPIIIa that differed only at amino acid residues 143. Whereas anti-Pena alloantibodies were able to recognize the Arg143 recombinant form of GPIIIa, anti-Penb antibodies were not. Conversely, anti-Penb alloantibodies were reactive only with the Gln143 isoform of the GPIIIa molecule. It thus appears that amino acid 143 of GPIIIa is not only associated with Pen phenotype, but specifically controls the formation and expression of the Pen alloantigenic determinants.

Enzymatic amplification of platelet-specific messenger RNA using the polymerase chain reaction.

Human platelets are derived from megakaryocytes as anucleate cells, and thus contain only vestigial amounts of RNA capable of being transcribed into protein. This has greatly hampered efforts to study directly platelet-specific gene products and their associated polymorphisms. In this report, we describe direct amplification, using the polymerase chain reaction, of platelet-derived mRNA in amounts sufficient to permit detailed analysis, such as restriction mapping and nucleotide sequencing. The ability to generate large amounts of cDNA from platelet-specific mRNA sequences should make possible direct molecular characterization of normal platelet proteins, and facilitate the investigation of a wide variety of inherited platelet disorders.

Human platelets circulating in mice: applications for interrogating platelet function and survival, the efficacy of antiplatelet therapeutics, and the molecular basis of platelet immunological disorders.

Herein we describe a novel animal model for examining the survival and function of human platelets following their circulation in non-obese diabetic/severe combined immunodeficient mice. Resting human platelets in platelet-rich plasma are introduced into the retro-orbital plexus, where they are absorbed with high efficiency and circulate for up to 2 days, comprising 10-20% of total circulating platelets. During this period of time, the human platelets can be exposed to a number of biochemical and immunochemical reagents, including novel antithrombotic compounds, or human antiplatelet antibodies that have been implicated in platelet destruction, activation or clearance. Platelets can also be subjected to a variety of storage conditions before infusion, and their relative survival and function following storage and circulation compared. The ability to evaluate in living mice the in vivo function and survival of circulating human platelets may prove valuable for determining mechanisms of antibody-mediated platelet passivation, and aid in the development of novel antiplatelet therapeutics.

Integrin clustering mechanisms explored with a soluble alphaIIbbeta3 ectodomain construct.

The purpose of this study was to test the hypothesis that residues critical for ligand- and temperature-induced clustering of integrin alphaIIbbeta3 are present on its extracellular domain. Sucrose density gradient sedimentation was used to examine the effects of ligand-mimetic peptides and physiological temperature on the oligomeric state of a soluble recombinant ectodomain variant of the alphaIIbbeta3 integrin, alphaIIbDelta962beta3Delta692, and its full-length counterpart. Both the ectodomain construct, isolated from High Five insect cell culture supernatants, and alphaIIbbeta3, isolated from human blood platelets, exhibited similar weight-average sedimentation coefficients at 23 degrees C, in the absence and presence of the ligand-mimetic peptide eptifibatide. These observations indicate that alphaIIbbeta3's ectodomain exhibits a similar extended conformation in both its free and ligand-bound states. Oligomerization was examined by incubation of both alphaIIbDelta962beta3Delta692 and full-length receptors at 37 degrees C, in the presence or absence of ligand-mimetic. Minimal oligomerization was observed with alphaIIbDelta962beta3Delta692. In contrast, full-length alphaIIbbeta3 exhibited substantial temperature-induced increases in its distribution of sedimenting species, indicative of thermal aggregation. These observations suggest that optimum oligomerization requires the participation of the integrin's transmembrane and cytoplasmic regions. In vivo, clustering of ligand-bound integrins may enhance signaling by increasing the local concentration of intracellular integrin-associated proteins.

Quinine-induced thrombocytopenia following intravenous use of heroin.

Profound thrombocytopenia developed in a 22-year-old man after intravenous use of heroin. A high-titer, quinine-dependent, platelet-specific antibody was detected in his serum using lysis of normal platelets labeled with chromium 51 and an electroimmunoassay for measurement of platelet-associated IgG. The antibody was specific for quinine and failed to react with platelets in the presence of quinidine hydrochloride or two structural analogues of heroin. Quinine, a common adulterant found in heroin, was detected in the patient's blood and urine. On the basis of these observations, the patient was judged to have quinine-induced immunologic thrombocytopenia. To our knowledge, this report is the first to confirm that quinine used as an adulterant can induce immunologic thrombocytopenia following an injection of heroin.

Can drugs cause autoimmune thrombocytopenic purpura?

A wide range of medications can cause life-threatening immune thrombocytopenia (ITP), hemolytic anemia, or neutropenia in sensitive individuals. The antibodies associated with these conditions usually require soluble drug to be present in order to react with the cell membrane glycoproteins for which they are specific. However, some patients make drug-independent antibodies (autoantibodies) as well. Occasionally, only autoantibodies are produced following exposure to a drug. Although drugs and other small molecules can become conjugated to proteins in vivo, which may induce an immune response, only fragmentary information is available to explain how exogenous substances sometimes perturb the immune system in such a way that antibodies capable of causing immune cytopenia are produced. Platelets are affected by drug-induced antibodies more often than any other blood element. For many drug-induced thrombocytopenias, the targeted membrane glycoproteins are readily accessible for laboratory investigation and methods for detecting the responsible antibodies are well developed. Techniques for studying cellular aspects of the immune response induced by drugs through in vitro manipulation of T and B lymphocytes are also advancing rapidly. Studies of drug-induced ITP may provide clues to the general mechanisms whereby drugs and other xenobiotics induce immune diseases. Clinicians should consider the possibility of an exogenous trigger in patients who present with apparent autoimmune thrombocytopenia.

Drug-induced immune thrombocytopenia: an overview of pathogenesis.

Many drugs are capable of causing antibody-mediated thrombocytopenia. Four, and perhaps five, different mechanisms have been implicated in the pathogenesis of this family of disorders. Some drugs become bound covalently to platelet membrane glycoproteins in vivo and stimulate the production of hapten-dependent antibodies that recognize drug-membrane protein targets. Others, such as quinidine, quinine, and sulfonamide antibiotics, induce the formation of an unusual class of antibodies that bind to membrane glycoproteins only when the drug (or one of its metabolites) is present in solution. Certain drugs trigger the production of true autoantibodies capable of binding to cell membrane glycoproteins in the absence of drug. Heparin-induced immune thrombocytopenia (HIT) is associated with antibodies specific for complexes formed between heparin and platelet factor 4 (PF4), a basic protein of the chemokine family found normally in platelet alpha granules. Immune complexes consisting of heparin, PF4, and antibodies are important in the pathogenesis of HIT, but the exact mechanisms by which they cause platelet destruction and, in some patients, thrombosis are not yet fully understood. Finally, thrombocytopenia in patients treated with recently introduced inhibitors of the platelet fibrinogen receptor (ligand mimetics) is thought to result from antibodies specific for ligand-induced binding sites (LIBS) on this receptor, but this mechanism has not yet been established.

Delayed thrombocytopenia after treatment with abciximab: a distinct clinical entity associated with the immune response to the drug.

BACKGROUND: Acute thrombocytopenia is a recognized side-effect of treatment with the fibrinogen receptor antagonist, abciximab, a chimeric (human/mouse) Fab fragment. The etiology of this complication is not fully understood. Generally, abciximab-induced thrombocytopenia occurs within a few hours of starting treatment with the drug. We have characterized a group of 13 patients who first developed thrombocytopenia 3-6 days after abciximab was discontinued. OBJECTIVE: To characterize clinical and serological aspects of this newly recognized clinical entity. PATIENTS AND METHODS: Clinical information was obtained from attending physicians and review of hospital records. Antibodies reactive with abciximab-coated platelets were characterized by flow cytometry. RESULTS: In each patient, IgG and/or IgM antibodies reactive with abciximab-coated platelets were identified. These antibodies could be distinguished from similar antibodies present in many normal persons by two criteria-they were relatively resistant to inhibition by normal Fab fragments, and they reacted preferentially with platelets coated with 7E3, the murine monoclonal antibody from which peptide sequences in abciximab are derived. Antibodies with these characteristics were not found in pretreatment serum from three of the thrombocytopenic patients or in patients given abciximab who did not develop thrombocytopenia. CONCLUSIONS: 'Delayed thrombocytopenia' after treatment with abciximab is caused by antibodies produced in response to the drug. These antibodies may be specific for murine peptide sequences in abciximab but could recognize other target epitopes on abciximab-coated platelets. Physicians administering abciximab should be aware of this potential complication of treatment, which usually occurs after discharge from hospital.

Persistence and progression of immunologic abnormalities in haemophilia.

Immunologic parameters in 51 patients with haemophilia treated with cryoprecipitate from volunteer donors or with commercial lyophilized concentrate were studied on 2-7 occasions. Concentrate users had significantly decreased mean T4 (helper)/T8 (suppressor) lymphocyte ratios when the initial (1.5 +/- .43) and most recent (1.25 +/- .60) values were compared (p less than .001), whereas in cryoprecipitate users mean values remained normal. Of the 22 concentrate users with initially normal T4/T8 ratios, 10 became abnormal during the study, an incidence of 45%. Fifteen of 28 (54%) patients with abnormal T4/T8 ratios had elevated circulating immune complexes (CIC), whereas all 16 patients with normal T4/T8 ratios had normal levels of CIC. Six patients have persistent lymphadenopathy.