Portal de Periódicos da CAPES

The dysfibrinogenaemias

2001; Wiley; Volume: 114; Issue: 2 Linguagem: Inglês

10.1046/j.1365-2141.2001.02892.x

1365-2141

Harold R. Roberts, Thomas E. Stinchcombe, Don A. Gabriel,

Iron Metabolism and Disorders

The dysfibrinogenaemias are characterized by structural abnormalities in the fibrinogen molecule that result in altered functional properties of the protein. These abnormalities are most often the result of inherited congenital abnormalities, and they will be the focus of this review. However, dysfibrinogenaemias can also be acquired as a result of underlying hepatic disease. Studies of abnormal fibrinogens are interesting not only from the standpoint of the clinical implications, but also because determination of the structural abnormalities that result in the dysfibrinogenaemias can be especially helpful in understanding the normal process of fibrin clot formation. Clinically, patients with inherited dysfibrinogenaemia are frequently asymptomatic; however, some patients will exhibit bleeding, thromboembolic phenomena or both. Asymptomatic patients with dysfibrinogenaemia are usually discovered incidentally because of abnormal coagulation screening tests. Most asymptomatic patients are heterozygous so that 50% of the fibrinogen molecules are normal, which is sufficient to maintain normal blood coagulation. Sometimes, however, heterozygotes will be symptomatic because the dysfibrinogen will interfere with one or more functional components of the normal fibrinogen molecule. A compilation of over 260 reported cases of dysfibrinogenaemia revealed that approximately 55% of the patients had no clinical manifestations; approximately 25% of cases exhibited a bleeding tendency; and 20% had a tendency towards thrombosis. Some patients who exhibited thrombosis also experienced bleeding symptoms (Ebert, 1994). The inherited dysfibrinogenaemias have traditionally been named after the location of their discovery or after the place of residency of the patient. With the advent of modern molecular biological techniques, a number of specific molecular abnormalities in the fibrinogen molecule have been identified. As a result, many of the dysfibrinogenaemias, previously thought to be distinct, have been found to have identical molecular defects. The acquired dysfibrinogenaemias are most often associated with underlying liver disease but clinical symptoms directly attributable to the acquired abnormality are uncommon, and most of the literature consists of case reports. To better understand the clinical consequences of the dysfibrinogenaemias, a brief review of the molecular biology and biochemistry of fibrinogen as well as the normal sequence of fibrinogen conversion to fibrin is needed. Fibrinogen is a 340 000-Da protein that is synthesized in hepatocytes (Forman & Barnhart, 1964; Fuller, 1983; Lord, 1995). The concentration of circulating fibrinogen is 1·5–3·5 g/l. Once secreted from the hepatocyte, the protein has a half-life of approximately 4 d and a catabolic rate of approximately 25% of the plasma pool per day (Collen, 1972). The molecule is a homodimer; each half consists of three non-identical polypeptide chains that have been termed Aα, Bβ, and γ chains (Fig 1). The genes for all three chains have been localized to the long arm of chromosome 4 (Henry et al, 1984). The amino termini of the three pairs of polypeptide chains are symmetrically arranged into a disulphide knot to form a central E-domain that is flanked by two large D-domains to form a trinodular structure (Fig 1) (Henschen, 1983; Doolittle, 1984; Huang, 1993; Everse et al, 1998). A schematic representation of the fibrinogen molecule. NH2 and COOH denote the amino and carboxy termini of the α chain (red), β chain (blue) and γ chain (green) respectively. SS identifies disulphide bonds. The E-domain is located in the central region of the molecule and is roughly composed of amino acids 1–49 of the α chain, 1–80 of the β chain, and 1–23 of the γ chain. The E-domain is flanked by two D-domains roughly composed of amino acids 111–197 of the α chain, 134–461 of the β chain, and 88–406 of the γ chain. The conversion of soluble fibrinogen into insoluble fibrin essentially consists of three main steps: release of fibrinopeptides A and B from the α and β chains, respectively, to form fibrin monomers; polymerization of the fibrin monomers to form a visible fibrin gel; and stabilization of the fibrin gel by activated factor XIII (Fig 2). Dysfibrinogens resulting in mutations that alter one or more of these steps have been identified. Fibrin formation is initiated when thrombin cleaves an arginyl residue at position 16 on the Aα chain releasing fibrinopeptide A (fpA) from each monomeric unit (Fig 2). This is followed by release of fibrinopeptide B (fpB) from the amino-terminal of the Bβ chain by thrombin cleavage of an arginyl residue at position 14 (Binnie & Lord, 1993). FpA is released initially at a more rapid rate than fpB, and is essential for the subsequent release of fpB (Blomback et al, 1978). A schematic representation of the conversion of fibrinogen to a fibrin clot by thrombin. Thrombin cleavage of fibrinopeptide A (amino acids 1–16 of the Aα-chain) followed by cleavage of fibrinopeptide B (amino acids 1–14 of the Bβ-chain) leads to formation of fibrin monomers. Fibrin monomers then self-assemble in a half-staggered linear overlap to form protofibrils. Protofibril formation is governed by the interaction of the thrombin exposed 'a' site on the E-domain with 'A' site on the D-domain. The protofibrils are then bundled into fibrin fibres through interaction of the 'b' site on the E-domain with the 'B' site located on the D-domain (see Fig 3). The release of fpA exposes binding sites in the E-domain, labelled 'a', which interact with binding sites on the D-domain, labelled 'A', of another fibrinogen molecule as shown in detail in Fig 3 (Blomback et al, 1978). There is a second site that is exposed on the E-domain after release of fpB, termed the 'b' site, which interacts with a corresponding 'B' site in the D-domain of another fibrinogen molecule (Fig 4). Schema showing the interaction of 'a' site on the E-domain of one fibrin monomer with the 'A' site on the D-domain of an adjacent fibrin monomer. This interaction contributes to the stabilization of protofibril formation. A schematic representation of the interaction of a 'b' site on one protofibril with a 'B' site on an adjacent protofibril. This interaction is thought to stabilize fibrin fibre formation. The exact role that each of these interactions plays in fibrin assembly has been explored using snake venoms that selectively cleave fpA and fpB (Li et al, 1996). Ancrod (Arvin) is purified venom of the Malayan pit viper, Agkistrodon rhodostoma, which selectively cleaves fibrinopeptide A. The enzyme batroxobin (Reptilase) from the South American pit viper, Bothrops atrox, is another snake venom that selectively cleaves only the A peptides. In experiments in which Reptilase is used, fibrin formation occurs, and it is morphologically similar to clots formed with thrombin (Weisel, 1986). However, when experiments are carried out using the enzyme from the copperhead snake, Ancistrodon contortrix, which preferentially cleaves fpB, no clot is formed at 37°C (Dardik & Shainoff, 1979; Shainoff & Dardik, 1983). Thus, it appears that the release of fpA is necessary for initial clot formation, whereas the release of fpB accelerates lateral aggregation and polymer formation theoretically by exposure of the 'b' binding sites and acceleration of the cross-linking of polymers by Factor XIIIa (Blomback et al, 1978; Weisel, 1986). The 'a' sites are believed to be on the γ chain contained within the E-domain and interact with a site within the D-domain of the γ chain of the adjacent fibrinogen molecule (Mosesson et al, 1995). There is evidence that a 'b' binding site occurs on the β chain and interacts with the 'B' site on the β chain on another protofibril in a similar fashion (Fig 4) (Everse et al, 1998). Factor XIII is converted to the active form, XIIIa, by thrombin. The rate of thrombin conversion from factor XIII to XIIIa is accelerated by fibrin polymerization (Greenberg et al, 1985). The presence of calcium and factor XIIIa will accelerate the cross-linking between the adjacent fibrin fibrils, creating an insoluble fibrin clot. Fibrinogen also plays a significant role in platelet aggregation and adhesion. The carboxyl-terminus of the γ chain in the region of amino acids 408–411 is believed to play a major role in the platelet adhesion (Liu et al, 1998). This region on the γ chain interacts with the activated glycoprotein IIb/IIIa receptor on platelets. At this time, no molecular defects in the γ chain region of 408–411 have been identified. Limited information exists about the binding sites that stabilize the protofibril–protofibril interaction. However, factors such as calcium, thrombin and fibrinogen concentrations as well as pH and fibre size have all been shown to alter the consistency and stability of the fibrin matrix. The thickness of the fibres influences the rate at which fibrinolysis occurs. The thinner fibres have been associated with a decreased rate of conversion of plasminogen to plasmin by tissue plasminogen activator (tPA) and are lysed more slowly than thick fibres (Gabriel et al, 1992). Molecular defects that result in dysfibrinogenaemia are usually caused by a single base mutation that results in the substitution of a single amino acid. Other dysfibrinogenaemias are the result of stop codons (resulting in a truncated molecule) or small base deletions or additions that alter normal fibrinogen structure. A database of known molecular defects can be found on the World Wide Web at http://www.geht.org/pages/database_fibrino_uk.html, entitled 'Fibrinogen Variants'. As might be expected, these structural modifications result in alterations in fibrinopeptide release, fibrin polymerization, fibrin cross-linking or fibrinolysis of the fibrin gel. In the homozygous form of dysfibrinogenaemia, the mutant chain appears in all fibrinogen molecules to form an abnormal homodimeric molecule. Heterozygotes are more complicated. For example, if there is a heterozygous mutation in the α chain, the resulting dimeric fibrinogen molecules may consist of homodimers (either normal, with both halves of the dimer containing the product of the normal allele, or abnormal, with both halves of the molecules containing the product of the abnormal allele). On the other hand, heterodimers with one-half of the dimeric fibrinogen containing the abnormal α chain and the other half containing the normal α chain may be present. Only a few of the patients with dysfibrinogenaemia are homozygotes. In the following discussion, patients are not defined in terms of whether they are hetero- or homozygotes, but rather whether they are symptomatic or asymptomatic. Although most homozygotes are symptomatic, some are not; and while many heterozygotes are asymptomatic, others are not (McDonagh, 2000). Illustrative defects of fibrinogen are depicted in Table I. The substitution of histidine for an arginyl residue at position 16 (Arg16→His) is responsible for a dysfibrinogen, previously termed fibrinogens Bern IV and Milano XI. This defect causes defective release of fpA and a delay in fibrin formation. Patients with this abnormality were found to be asymptomatic or to exhibit a mild bleeding disorder (Stucki et al, 1999). The dysfibrinogenaemia known as fibrinogen Bremen illustrates the critical role that spatial arrangements play in the polymerization of the protofibrils. In this dysfibrinogen, the glycine 17 residue, near the amino terminus of the α chain, is replaced with a valine (Gly17→Val). This substitution is characterized by delayed polymerization as a result of decreased affinity between the complementary 'A-a' sites, thus inhibiting the D–E interactions necessary for polymerization (Wada et al, 1993). The proband for this dysfibrinogen had a history of easy bruising, subcutaneous haematomas and delayed postoperative wound healing. An Arg554→Cys substitution on the Aα chain (previously named Chapel Hill III, Paris V and Dusart) predisposes patients to thrombosis (Wada & Lord, 1994). The cysteine residue covalently binds albumin to the C-terminus of the Aα chain. Albumin binding then induces formation of abnormally thin fibrin fibres resistant to plasmin degradation (Wada & Lord, 1994). Thus, it appears that the impaired fibrinolysis exhibited by this dysfibrinogen is responsible for the thrombotic manifestations observed in these patients (Haverkate & Samama, 1995). Another example of an α-chain mutation that results in abnormal albumin binding and plasmin resistance is dysfibrinogen Marburg, caused by an Aα→Lys461 base substitution leading into a premature stop codon and a truncated fibrinogen molecule. This leads to a cysteine residue not having its 'partner' to form a disulphide bridge. Thus, the unpaired cysteine binds to albumin. The consequence is resistance to plasmin digestion as well as to impaired lateral association of the protofibrils (Sugo et al, 1998). The clot formed by this dysfibrinogen is fragile and fails to form normal fibrin aggregates. The fibres that are formed are very thin. Thus, fibrin formation is inadequate and results in a bleeding tendency. However, there is also impaired fibrinolysis, leading to thrombotic complications. The proband with dysfibrinogen Marburg experienced postoperative pelvic vein thrombosis and recurrent pulmonary embolism as well as severe uterine bleeding following a Caesarean section. The dysfibrinogen Bern V is characterized by an Arg16→Cys mutation which has been shown to result in abnormal albumin binding and delayed fibrinolysis as well (Stucki et al, 1999). Of the patients reported with this amino acid substitution, some have been asymptomatic, others have experienced excessive postpartum bleeding, while others have had thrombotic events such as spontaneous abortions (Ebert, 1994). This example typifies the variability in clinical manifestations of some dysfibrinogens. Another defect that has unusual pathophysiology is dysfibrinogen Lima, characterized by an Arg141→Ser substitution in the Aα chain. As a result, there is abnormal glycosylation of an aspartic acid residue at 139 (Maekawa et al, 1992). The abnormal sialylation leads to an abnormally negatively charged molecule, which is believed to result in repulsive forces that contribute to a delay in the lateral association of fibrin fibres and a subsequent delay in fibrin gel formation. When the fibrinogen molecule undergoes desialylation, the fibrinogen polymerization improves to near normal. Despite the laboratory defect, however, patients with this dysfibrinogen had no clinical symptoms, even though one patient was homozygous for the disorder. Beta-chain mutations are less frequent than α- or γ-chain mutations. An example of defective polymerization is fibrinogen Ise (Fukuoka II) secondary to a Gly15→Cys mutation on the Bβ chain (Yoshida et al, 1991). This abnormality was found to result in delayed release of fpB, a prolonged thrombin time and defective polymerization of the fibrin monomers. The defect was not associated with any bleeding in the propositus or in three relatives discovered on family screening. A β-chain mutation, fibrinogen Naples (Ala68→Thr), results in defective binding of thrombin to fibrin and decreased release of fpA and fpB in homozygous and heterozygous individuals (Koopman et al, 1992). Fibrin from homozygous individuals bound approximately 10% of the thrombin compared with normal fibrin, whereas fibrin from heterozygous individuals bound approximately 50% of the thrombin. Interestingly, in the three family members who were homozygous for fibrinogen Naples, there was a history of arterial thrombosis occurring at 21 and 25 years of age, and of deep vein thrombosis occurring at age 33 years. The mechanism responsible for the thrombotic manifestation in the homozygous patients is not clear-cut. However, it is postulated that thrombin binding to fibrin is essential to prevent thrombosis in large vessels in which there is less thrombomodulin and less inhibition by antithrombin than in microcirculation. Interestingly, the three heterozygous members of the family were completely asymptomatic. Dysfibrinogen New York I is caused by a deletion of amino acids 9–72, which corresponds to exon 2 of the β-chain gene (Liu et al, 1985). This deletion results in the absence of the cysteine residue at amino acid 65 of the β chain that normally forms a disulphide bond with a cysteine residue on the α chain. The absence of the disulphide bond is believed to result in a conformational change in the amino-terminal region of the molecule, and results in a delayed release of fpA and fpB. Samples from the two heterozygous propositi revealed that each patient had 35–50% non-clottable fibrinogen compared with 2% non-clottable fibrinogen in normal controls. The patient who was initially identified with this disorder had multiple thrombotic events, which led to her death. Of two brothers with thrombotic tendencies, one died from a pulmonary embolism 6 weeks after a cholecystectomy, and the other brother experienced non-fatal thrombotic events. Nine siblings were unaffected. Interestingly, her parents' deaths were not related to thrombosis, and they had no prior history of thrombosis (Al-Mondhiry et al, 1975). Of the known gamma-chain mutations, approximately 5% of patients had a clinical history of significant bleeding, while about 30% had thrombotic tendencies. The remaining 65% of patients were asymptomatic (Côtéet al, 1998). An example of a γ-chain mutation associated with bleeding is a patient with fibrinogen Asahi (Met310→Thr). The propositus was a 33-year-old man who, since adolescence, had mild to moderate bleeding after injuries (Yamazumi et al, 1989). The Met310→Thr mutation impaired glycosylation of the asparagine residue at position 308. This substitution resulted in impaired fibrin polymerization and delayed cross-linking of the fibrin gel. Presumably the extra oligosaccharide altered the conformation of the d-domain in a manner that impaired polymerization (Yamazumi et al, 1989). A dysfibrinogen Asn308→Lys has reportedly resulted in both bleeding and thrombosis in different individuals (Grailhe et al, 1993). Of the γ-chain mutations associated with thrombosis, five involved substitution of the usual arginyl residue at position 275. Two of them (Bologna I, Cedar Rapids I) had an Arg275→Cys substitution, whereas three (Barcelona III, Bergamo II and Haifa I) were caused by Arg275→His. The three propositi of fibrinogen Cedar Rapids all had pregnancy-associated thrombosis and were all heterozygous for Factor V Leiden as well the dysfibrinogenaemia (Mosesson, 1999). Of note, only the patients who were heterozygous for both the Factor V Leiden and the Arg275→Cys substitution were symptomatic, whereas other family members with only one of the two abnormalities had no clinical symptoms. This suggests that both defects were synergistic in causing thrombosis. Recently, dysfibrinogen Bellingham, also caused by an Arg275→Cys substitution, was reported in a man who had multiple thromboembolic events (Linenberger et al, 2000). However, this patient also had a G/A transition at position 455 in the β-chain promoter region of the gene. The combined genotype is associated with high acute-phase levels of fibrinogen. The seven family members who were heterozygous for the Arg275→Cys had no thromboembolic events, and the three family members who only had the isolated β-chain promoter defect also had no history of thrombosis, again suggesting that the Arg275→Cys substitution alone is asymptomatic but results in thrombosis when combined with another defect. This set of disorders has been associated with advanced liver disease, including cirrhosis and hepatocellular carcinoma. In the setting of liver disease, dysfibrinogenaemia is believed to develop as a result of an abnormal increase in the number of sialic acid residues, causing impaired fibrin monomer polymerization (Martinez et al, 1983). When excess sialic acids are removed, normalization of the thrombin time and fibrin monomer aggregation occurs. The dysfibrinogenaemias associated with hepatoma are also known to have an increased sialic acid content (Gralnick et al, 1978). Acquired abnormalities mimicking dysfibrinogenaemias include multiple myeloma and Waldenstrom's macroglobulinaemia. In these patients, impaired polymerization of fibrin monomers occurs as a result of interference of polymerization of the normal monomers by a paraprotein (Coleman et al, 1972; Gabriel et al, 1983). Autoimmune diseases may mimic dysfibrinogenaemias because some acquired antibodies inhibit fibrin polymerization or delay fibrinopeptide release (Galanakis et al, 1978; Marciniak & Greenwood, 1979). Hydroxyethyl starch and dextran are used as colloid volume replacement in some surgical patients, and these macromolecules can alter the fibrinogen levels in patients. Because both agents can cause significant dilutional effects, the fibrinogen levels may be artificially low. If the samples are corrected for the increased plasma volume the fibrinogen measurement of patient samples after the use of these agents may be artificially elevated (Hiipala, 1995). This is believed to be caused by the fact that the dextran and hydroethyl starch cause increased light scatter in the patient samples; therefore, indirect fibrinogen assays that rely on this method can result in an artificially elevated level. The thrombin and Reptilase clotting times are shortened after the use of these agents, probably because of accelerated clot formation or changes in the photo-optical properties of the sample. The urokinase-activated clot lysis times are also shortened, which suggest accelerated clot lysis (Strauss et al, 1985). Patients with asymptomatic dysfibrinogenaemias are usually identified by prolongation of routine screening tests of coagulation: thrombin clotting time (TCT); prothrombin time (PT); and partial thromboplastin time (PTT). Often the abnormalities are discovered incidentally in patients scheduled for surgery or other invasive procedures. Occasionally, a patient will present with clinical signs of excessive bleeding and/or wound dehiscence after a minor operation. If the diagnosis is clinically suspected, the initial work-up should consist of TCT, PT and PTT. Although all three tests are usually prolonged, the TCT is the most sensitive of the three. An algorithm for the diagnosis of dysfibrinogenaemia based on the TCT as an initial test is shown in Figure 5. The TCT will also be prolonged in the presence of heparin, hypofibrinogenaemia, afibrinogenaemia, dysfibrinogenaemia or in disseminated intravascular coagulation DIC), when there are excessive fibrin split products. In addition to these screening tests, a Reptilase or Ancrod time should also be ordered. These tests use a snake venom that directly cleaves the A peptide from the Aα chain of fibrinogen at an identical site as thrombin. The Reptilase and Ancrod tests are not affected by the presence of heparin or antithrombin III. Therefore, the times will be within normal limits if the prolonged TCT is caused by contamination with heparin, but prolonged in the presence of a dysfibrinogen. In some patients with dysfibrinogens predisposing to thrombosis, the TCT may be significantly shorter than normal, as in fibrinogen Oslo I. Turbidity kinetics and light-scattering technology (Carr & Hermans, 1978) can also be used to detect abnormal fibrinogen molecules (Carr et al, 1976). Diagnostic algorithm for dysfibrinogenaemia. *Frequently normal, but can be occasionally low. **Almost always low, but can be normal. The activity may be altered if fibrin degradation products are elevated. Qualitative fibrinogen levels should be measured by determining the amount of clottable fibrinogen using a functional assay and the amount of fibrinogen protein as determined by immunological methods. In some dysfibrinogenaemias, fibrinogen functional activity will be low, whereas the fibrinogen antigen levels will be normal. However, in other dysfibrinogens there is concordance between the functional and immunological concentrations. Fibrinogen levels are often elevated in chronic inflammatory states such as collagen vascular diseases, and can be elevated after trauma or serious medical illness owing to its role as an acute-phase reactant. Although these clinical situations may cause increased levels of fibrinogen, they will not change the impaired functional activity of a dysfibrinogen on a functional assay, or the altered TCT or Reptilase time. The gold standard for the diagnosis of a dysfibrinogenaemia is identification of the molecular defect. Because most clinical laboratories do not do these tests, referral to a reference laboratory may be required. It is important to determine whether the patient is heterozygous or homozygous for the dysfibrinogen as this can play a critical role in clinical decision-making because homozygous patients tend to have more severe symptoms. Another important aspect of the work-up is to evaluate family members for the potential defect, not only to provide insight into the clinical behaviour of the dysfibrinogen but also for genetic counselling. The prevalence of dysfibrinogenaemias in patients with deep-vein thrombosis is 0·8% based on a review of nine studies involving 2376 patients in seven different countries (Haverkate & Samama, 1995). When the dysfibrinogen is diagnosed in the setting of an idiopathic deep-vein thrombosis, it is important to seek additional factors that may have predisposed to the event, including the use of oral contraceptives, long car or plane rides or prolonged bed rest as a result of a medical or surgical illness. An evaluation for a co-existing genetic causes of thrombophilia should also be undertaken, including evaluation for defects in the protein C system, the prothrombin mutation, factor V Leiden and other known predisposing risk factors. The patients with dysfibrinogenaemia associated with bleeding tendencies often present with bleeding after surgery or trauma. Many women are first diagnosed after having prolonged postpartum bleeding. Occasionally, a patient will have difficulty with wound healing or experience wound dehiscence. Some patients report a history of easy bruising or bleeding after minor cuts or after trauma. The dysfibrinogenaemias most often associated with bleeding disorders are mutations on the α chain, especially in the amino-terminal region (amino acids 17–19). This region is believed to be a critical site for polymerization between adjacent fibrin monomers. Some of these mutations have also been associated with delayed fibrinopeptide release as well. Although many people with a mutation in this region exhibit a bleeding tendency, some members in the same family with the same mutation are asymptomatic. The rarity of the disorder and the tremendous variability in the clinical course make the clinical management of these patients difficult. Although general guidelines concerning management of patients with dysfibrinogenaemias can be formulated, each patient will need to be evaluated on an individual basis. In general, if the patient has a dysfibrinogenaemia with a known history of previous bleeding, fibrinogen replacement should be given prior to surgery or after trauma. Fibrinogen replacement should be sufficient to raise the level of fibrinogen to haemostatic levels (approximately 1 g/l, using cryoprecipitate as a source of fibrinogen). Each bag of cryoprecipitate will raise the fibrinogen level about 0·1 g/l with a half-life of 2–4 d. Care should be taken to ensure that the source of cryoprecipitate adheres to accepted safety standards. In some areas, solvent detergent-treated cryoprecipitate is available. In other areas, purified fibrinogen free of transmissible agents can be obtained. If the patient has tolerated invasive procedures and operations in the past without bleeding, it is likely that further invasive procedures can be undertaken safely without replacement therapy, although close observation of the patient is essential. Antifibrinolytic amino acids such as aminocaproic acid and tranexemic acid can be used in dysfibrinogenaemia patients who have bleeding symptoms, but such agents are not recommended for patients with dysfibrinogens associated with thrombosis. Currently it is believed that there are two mechanisms predominantly responsible for the episodes of thrombosis associated with dysfibrinogenaemias. The first is that the abnormal fibrinogen is defective in binding thrombin, which results in elevated levels of thrombin. The second is that the abnormal fibrinogen forms a fibrin clot that is resistant to plasmin degradation. In 1995, the Scientific and Standardization Committee (SCC) of the International Society on Thrombosis and Haemostasis reviewed 51 cases of dysfibrinogenaemia associated with thrombophilia (Haverkate & Samama, 1995). Twenty-six cases fulfilled the criteria set up by the committee for familial dysfibrinogenaemia and for thrombosis that was not caused by other factors. In many of the cases, the proband had family members who also experienced thrombosis. In this series, 7 out of 15 of the women who had pregnancies developed thrombosis after delivery. There was also an increased number of spontaneous abortions and stillbirths. The ages of the patients with confirmed dysfibrinogenaemia associated with thrombophilia ranged from 16 to 50 years. Many of the patients who had suffered a thrombotic episode also had postpartum bleeding or a history of mild bleeding. If the patient has not had a documented thrombosis and is found to have a dysfibrinogenaemia with a molecular defect associated with thrombosis, continued observation is recommended in view of the tremendous variability in the clinical behaviour of these disorders. Patients should be encouraged to take such precautionary measures as avoidance of venous stasis and early ambulation after any surgical intervention. At this time there is not enough data to recommend routine anticoagulation during pregnancy for patients without a previous episode of documented thrombosis. Furthermore, any potential benefit from anticoagulation must weighed against the potential risk of bleeding complications that may be higher in this patient population as many also have a history of a mild bleeding diathesis. Once a patient has a documented thrombotic episode, the patient should receive anticoagulation therapy, initially with heparin and later with warfarin, unless there is a contraindication to anticoagulation. The optimal duration of anticoagulation for patients with thrombosis caused by dysfibrinogenaemia has not been clearly established. Most often the decision will depend on the clinical situation, potential contributing factors to the thrombosis, the particular type dysfibrinogenaemia and the clinical course of the dysfibrinogenaemia in other family members. If all secondary causes of thrombosis have been eliminated and it appears that an isolated dysfibrinogenaemia is the cause of the thrombosis, then the determination of the specific dysfibrinogen and whether the patient is homozygous or heterozygous is indicated. If the patient has had multiple thromboembolic events, a spontaneous life-threatening event (i.e. near-fatal pulmonary embolus, mesenteric thrombosis or portal-vein thrombosis), or has more than one genetic defect predisposing to thrombosis, indefinite anticoagulation has been recommended by some authorities. If the patient has had only one thrombotic event with a known provocative stimulus, anticoagulation for a period of 6 months may be sufficient, provided the patient is aware of the risks and of the symptoms of a recurrent thrombosis. Like all genetic disorders, studies of the dysfibrinogenaemias are illustrative of the genetic heterogeneity of the population. It also appears that most of the abnormal fibrinogens associated with bleeding are associated with mutations in the amino-terminal region of the Aα chain (Lord, 1995). Variants in the amino terminal of the Bβ and Aα chains can also be asymptomatic. Thrombotic manifestations are often associated with variants in which substitutions occur leading to disulphide linkage to albumin. These ultimately interfere with fibrinolysis of the clot by decreased binding of tissue plasminogen activator and/or formation of thin fibrin fibres resistant to lysis. Further studies of the dysfibrinogens promise to lead to further understanding of fibrin formation, haemostasis, and thrombosis.