Article in press - uncorrected proof

J Lab Med 2008;32(6):401–405 2008 by Walter de Gruyter • Berlin • New York. DOI 10.1515/JLM.2008.059 2008/59

Ha¨ mostaseologie Redaktion: C.M. Schambeck

Laboratory diagnosis and clinical manifestations of patients with dysfibrinogenemia1) Laboratoriumsdiagnostik und klinische Manifestation der Dysfribrinogena¨ mie

Wolfgang Miesbach* Keywords: ; dysfibrinogenemia; ; . Medical Clinic III/Institute of Transfusion Medicine, Goethe University Frankfurt, Frankfurt am Main, Zusammenfassung Germany Die angeborene Dysfibrinogena¨ mie ist eine sehr seltene Gerinnungssto¨ rung, basierend auf einem strukturellem Abstract Defekt im Fibrinogenmoleku¨ l, worunter eine Neigung zu Blutungen, Thrombosen und Aborten entstehen kann. Hereditary dysfibrinogenemia is a rare clotting disorder Wir beschreiben die Ergebnisse der Labordiagnostik und due to a structural defect in the molecule that die klinischen Manifestationen bei 50 Patienten mit Dys- results in a tendency for bleeding and thrombosis, as well fibrinogena¨ mie. Es wurden folgende Untersuchungen as obstetric complications. We describe the laboratory vorgenommen: Fibrinogen (Clauss), immunologisches results and clinical manifestations for 50 patients with a Fibrinogen und Hitze-Fibrinogen nach Schulz. Unter den diagnosis of dysfibrinogenemia. Various different labo- 50 Patienten mit Dysfibrinogena¨ mie (52% weiblich; ratory measurements of fibrinogen were performed on Altersmedian 52 Jahre, 9–89 Jahre) war die Fibrinogen- samples from these patients, including fibrinogen konzentration nach Clauss erniedrigt im Median mit (Clauss), heat fibrinogen precipitation according to 51 mg/dL (15–86 mg/dL; Norm 150–450 mg/dL). Die Schulz and immunological fibrinogen. Fifty patients were Bestimmungen der Fibrinogen-Antigen-Konzentrationen found with dysfibrinogenemia (52% female; median age ergaben Normalbefunde: Hitze-Fibrinogen nach Schulz, 52, range 9–89 years). The fibrinogen level according 240 mg/dL; und immunologisches Fibrinogen, 244 mg/dL. to Clauss was low, with a median of 51 mg/dL (range Die Reptilasezeit war mit 55 s verla¨ ngert (Median; Norm 15–86 mg/dL; normal range 150–450 mg/dL). Determi- -20 s). Fu¨ nfzig Prozent der Patienten hatten ein oder nation of other fibrinogen levels revealed normal results: mehrere Blutungsereignisse, meistens eine deutliche Nei- heat fibrinogen precipitation according to Schulz, gung zu Ha¨ matomen (60%) oder Nachblutung (44%). 240 mg/dL; and immunological fibrinogen, 244 mg/dL. Bei 13 Patienten ist ein thromboembolisches The median was longer than normal, at Ereignis aufgetreten. Bei 12% der Patienten war eine 55 s (normal 20 s). Some 50% of the patients reported Neigung gleichzeitig zu Blutung und Thrombosen vor- a distinct bleeding tendency, mostly a tendency for handen. Bei 19% der weiblichen Patienten hatten einen (60%) and secondary bleeding (44%). Thir- oder mehrere Aborte. Bei Patienten mit stattgehabten teen patients had thrombotic events, of which 54% were Blutungssymptomen war die Konzentration von Fibrino- located arterially. Some 12% of the patients repor- gen nach Clauss mit 43 mg/dL im Median niedriger als ted a tendency for bleeding and for thrombosis, whereas bei Patienten ohne Blutungsproblemen (57 mg/dL). 19% had , sometimes recurrent. We found that functional fibrinogen levels (Clauss) were generally lower in patients with bleeding manifestations (43 vs. Schlu¨ sselwo¨ rter: Abort; Blutung; Dysfibrinogena¨ mia; 57 mg/dL in other patients). Thrombose.

1)Zusammenarbeit von DGKL-Fachredaktion und GTH-Arbeits- gruppe MAISTHRO (Main-Isar-Thromboseregister). Overview *Correspondence: Wolfgang Miesbach, MD, Medical Clinic III/Institute of Transfusion Medicine, Goethe University Fibrinogen is a 340,000-Da protein that is synthesized in Frankfurt, Theodor-Stern-Kai 7, 60590 Frankfurt, Germany . Fibrinogen is the precursor of and is Tel.: q49-69-63015051 Fax: q49-69-63016738 thus a key component in clot formation. The develop- E-Mail: [email protected] ment of the three-dimensional network in fibrin can be Article in press - uncorrected proof

402 Miesbach: Diagnosis of patients with dysfibrinogenemia

divided into three steps w1x. First the fibrinogen molecule have shown that IgG antibodies directed against the is activated by , which removes two pairs of fibrinogen molecule interfere with the binding of thrombin small peptides from the Aa and Bb chains, creating two w18x and that monoclonal immunoglobulin g light chains polymerization sites known as A9 and B9 w2, 3x. The sec- bind to fibrinogen and interfere with all stages of poly- ond step is the formation of intermediate polymers. The merization w19x. Acquired dysfibrinogenemia may have a fibrin monomers associate in a half-staggered manner poorer prognosis than congenital dysfibrinogenemia through E and D regions and end-to-end through the D owing to possible underlying co-morbidities. A wide vari- regions of different monomer units, giving rise to longi- ety of symptoms from bleeding to thrombosis are tudinal growth of a polymer with a width of two fibrin observed in affected patients w16x. Besides some case monomers, yielding the fibrin monomer unit called pro- studies, there are only a few studies on the clinical tofibril. The third step is lateral association of these inter- manifestations in a large number of patients with mediate polymers, forming the three-dimensional fibrin dysfibrinogenemia. network w1x. The obstetric complications of dysfibrinogenemia The fibrinogen molecule is a homodimer; each half include miscarriage, mostly during the first trimester, consists of three non-identical polypeptide chains that thrombosis, bleeding and placental abruption during are termed Aa,Bb and g chains w4x. The genes for all pregnancy. There are very few studies on the impact of three chains have been localized to the long arm of chro- dysfibrinogenemia on pregnancy or on treatment mosome 4. The amino termini of the three pairs of poly- approaches. For patients with congenital dysfibrinogen- peptide chains are symmetrically arranged in a disulfide emia, effective treatment has proven extremely difficult, knot to form a central E-domain that is flanked by two since there is a broad spectrum of symptoms that are large D-domains to form a tri-nodular structure w5x. The sometimes contradictory. Our study was designed to D region is divided into several domains, including the b establish a correlation between variations in clinical and g nodules, which constitute the C-terminal parts of manifestations and laboratory test results. these chains that fold independently w6x. The C-terminal a-chain, spanning the sequence from Aa-208 to Aa-610, contains a nodule and a connecting strand w7x. The aC Patients and methods domains have several functions in the hemostatic pro- cess and act as substrates for activated factor XIII w8–12x. We investigated 50 patients from 19 families with dys- The aC domains contribute to the polymerization pro- fibrinogenemia, of whom 26 (52%) were female. The cess, accelerating it and influencing the final clot struc- median patient age was 52 years (range 9–89 years). ture w13, 14x. Amino acids Aa 572–575 are involved in Each patient was interviewed regarding history of clinical the adhesion of other proteins and structures such as manifestations. endothelial cells. Laboratory diagnosis of dysfibrinogenemia was based Congenital dysfibrinogenemia is caused by genetic on criteria established by the International Society of defects in fibrinogen, resulting in a decrease in functional Thrombosis and Hemostasis (Subcommittee on Fibrino- clotting activity despite normal circulating levels of fibrin- gen; Scientific and Standardization Committee) w20x. ogen antigen. Congenital dysfibrinogenemia may be Testing started with sensitive but non-specific screening transmitted in a homo- or heterozygous manner, and tests and then included more specific confirmatory tests. inheritance is mostly autosomal dominant or co-domi- Initially, all patients had to undergo standard screening nant. Most homozygous carriers are symptomatic, tests: , prothrombin time, and activated whereas some heterozygous carriers are asymptomatic partial thromboplastin time. Plasma concentrations of w15x. The most common structural defects involve the fibrinogen were analyzed for all patients using several fibrinopeptides and their cleaving sites; the second most methods. In most cases, laboratory diagnosis of dysfi- common involves the g-chain polymerization region. brinogenemia can be established from a discrepancy These defects are primarily the result of DNA point muta- between fibrinogen activity and fibrinogen antigen levels tions with substitutions of single amino acids. Other dys- w23x. fibrinogenemias are caused by formation of new stop Fibrinogen activity was measured in terms of fibrin codons, and small base additions or deletions. Currently, polymerization function by the Clauss method, which a database of human fibrinogen variants includes more measures the rate of clot formation after adding a high than 430 molecular abnormalities (http://www.geht.org/ concentration of thrombin to citrated plasma. Fibrinogen databaseang/fibrinogen). Dysfibrinogenemia may also be antigen concentrations were determined by immunologic acquired secondary to , the use of certain (radial immunodiffusion) w22, 23x and precipitation (heat drugs (e.g., acetaminophen) w16, 17x, autoimmune dis- according to Schulz) methods w23x. ease, and malignancies, and may even occur sponta- neously without a known underlying cause. There are case reports of patients presenting with Results newly acquired abnormal laboratory results, including prolonged thrombin and reptilase time and low All 50 patients showed abnormal laboratory results in functional fibrinogen (Clauss) levels w18, 19x. Studies terms of low functional fibrinogen levels (Clauss) in con- Article in press - uncorrected proof

Miesbach: Diagnosis of patients with dysfibrinogenemia 403

trast to normal values for heat precipitation fibrinogen Miscarriages occurred in five of 26 female patients (Schulz) and immunologic fibrinogen (Table 1). (19%). Three of these patients suffered from recurrent The median fibrinogen concentration (Clauss) was pregnancy loss. 51 mg/dL (range 15–86 mg/dL; normal range There were several additional laboratory findings in 150–450 mg/dL). The median immunological fibrinogen patients with dysfibrinogenemia. Von Willebrand syn- concentration was 244 mg/dL (normal range drome was present in one family, with a combination of 200–450 mg/dL) and the median concentration of heat bleeding and thrombotic features (two patients with precipitation fibrinogen (Schulz) was 240 mg/dL (normal bleeding symptoms, one patient with bleeding and DVT). range 200–450 mg/dL). The median reptilase time was Heterozygous factor V-Leiden was found in two 55 s (normal -20 s). patients from one family (one patient with MI and the We found that functional fibrinogen (Clauss) was other patient was asymptomatic). Heterozygous factor II generally lower in patients with bleeding manifestations mutation (prothrombin 2021 GA) was found in one family (43 mg/dL) compared to other patients (57 mg/dL). The (one patient with and gastrointestinal bleeding). clinical parameters for the patients are summarized in Table 2. Despite having abnormal laboratory results, 18/50 Discussion study participants showed no symptoms of dysfibrino- genemia. Various degrees of bleeding were described in The clinical manifestations in patients with dysfibrino- 25/50 patients (50%). Six of 50 patients (12%) manifest- genemia include bleeding symptoms, venous or arterial ed with both hemorrhage and thrombosis. The most thrombosis and abortions. It is not clear, however, which common hemorrhagic manifestation was a tendency factors precipitate the development of one of these for bruising/hematoma, found in 15/25 (60%) of patients symptoms in patients. All of our patients had low levels with hemorrhagic symptoms. This was followed by of functional fibrinogen but normal fibrinogen antigen excessive post-traumatic bleeding (including post-surgi- levels, which is significant for the laboratory diagnosis cal and post-partal) in 11/25 (44%) of the patients, oral of dysfibrinogenemia w21x. (dental, gums) bleeding in 6/25 (24%), epistaxis in 5/25 We compared the results of our study to a meta-analy- (20%), excessive menstrual bleeding in 4/25 (16%), and sis of over 260 reported cases of dysfibrinogenemia, gastro-intestinal bleeding in 2/25 (8%). which revealed that approximately 55% of patients had In 13/50 (26%) of the patients, thrombotic events no clinical manifestation; approximately 25% of patients occurred. The overall number of patients with deep exhibited a tendency for hemorrhage and 20% had a ten- (DVT) was six, with pulmonary embo- dency for thrombosis w20x. Interestingly, 27% of patients lism in 50% of these. Arterial thrombosis occurred in sev- with a history of thrombosis also experienced bleeding. en patients. In patients with arterial thrombosis events, Some 50% of our study patients suffered from bleeding, 28% suffered from (MI), 42% from mostly recurrent, spontaneous not caused ischemic stroke and 28% of the patients had peripheral by injury. These findings confirm the results of a previous artery disease. In two patients, both venous and arterial study that bleeding is the most frequent symptom in dys- thrombosis was found. fibrinogenemia w20x. We found that functional fibrinogen level (Clauss) was generally lower in patients with bleed- Table 1 Laboratory criteria for diagnosis of patients with ing manifestations. dysfibrinogenemia. The rate of clinical manifestations, particularly bleed- ing, was strikingly higher than in previous studies. This Parameter Dysfibrino- Normal might be explained by selection bias. Our study popu- genemia lation was recruited from outpatients attending our clinic Median Range and their family members. Fibrinogen (Clauss), mg/dL 51 10–86 150–450 In general, patients with a known history of previous Heat fibrinogen, mg/dL 240 200–450 bleeding should be treated with a fibrinogen substitute Immunological fibrinogen, mg/dL 244 200–450 prior to surgery, after trauma or post partum. Fibrinogen Reptilase time, s 55 -20 substitution needs to be sufficient to raise the plasma level of fibrinogen to hemostatic levels. In patients with Table 2 Clinical manifestations in patients with dysfibrino- thrombosis the prevalence of dysfibrinogenemia is rather genemia. low. In a study of 2376 patients with DVT the prevalence of dysfibrinogenemia was 0.08% w24x. In many cases, Clinical manifestation Patients (%) the patients had family members with a history of thrombosis. Asymptomatic 18/50 (36) Bleeding 25/50 (50) In patients with dysfibrinogenemia, thrombosis is usu- Thrombosis 13/50 (26) ally mild and arterial thrombosis is very rare. Surprisingly, Arterial thrombosis 9/13 (69) the majority of patients with thrombotic events in our Bleeding and thrombosis 6/50 (12) study suffered from arterial thrombosis (54%). Several Abortion 5/26 (19) mechanisms may be responsible for the occurrence of Article in press - uncorrected proof

404 Miesbach: Diagnosis of patients with dysfibrinogenemia

thrombosis in patients with dysfibrinogenemia w24, 25x. 3. Olexa SA, Budzynski AZ. Evidence for four different poly- One is defective binding of abnormal fibrinogen to throm- merization sites involved in human fibrin formation. Proc bin, which results in elevated levels of thrombin. Abnor- Natl Acad Sci USA 1980;77:1374–8. 4. Henry I, Uzan G, Weil D, Nicolas H, Kaplan JC, Margueric mal fibrinogen may also form a fibrin clot that is resistant C, et al. The genes coding for A alpha, B beta, and gamma to plasmin degradation. Moreover, decreased tissue chains of fibrinogen map to 4q2. Am J Hum Genet 1984; plasminogen activator binding and/or plasminogen bind- 36:760–8. ing in some dysfibrinogens may result in lower plasmin- 5. Martinez J. Congenital dysfibrinogenemia. Curr Opin Hema- ogen activation w26x. It has been suggested that altered tol 1997;4:357–65. clot architecture resulting in proteolytic resistance is 6. Weisel JW, Stauffacher CV, Bullitt E, Cohen C. A model for fibrinogen: domains and sequences. Science 1985;230: responsible for thrombosis w27x. 1388–91. We had a lower number of asymptomatic patients in 7. Spraggon G, Everse SJ, Doolittle RF. Crystal structures of our population, most likely owing to the selection bias fragment D from human fibrinogen and its cross-linked already noted for patients with bleeding symptoms. counterpart from fibrin. Nature (London) 1997;389:455–62. Moreover, investigation of fibrinogen gene was 8. Cottrell BA, Strong DD, Watt KWK, Doolittle RF. not within the scope of our study. Therefore, we cannot studies on the a chain human fibrinogen. Exact location of relate the variability of manifestations to underlying gene cross-linking acceptor sites. Biochemistry 1979;18:5405– defects. 10. 9. Sobel JH, Gawinowicz MA. Identification of the a chain Our results demonstrate the variability of clinical man- lysine donor sites involved in factor XIIIa fibrin cross-linking. ifestations in families with congenital dysfibrinogenemia. J Biol Chem 1996;271:19288–97. The distribution of manifestations within families is as 10. Roberts WW, Lorand L, Mockros LF. Viscoelastic properties variable as the disease itself. This variability of clinical of fibrin clots. Biorheology 1973;10:29–42. manifestations of dysfibrinogenemia may reflect the 11. Shen L, Lorand L. Contribution of fibrin stabilization to clot genetic heterogeneity of the syndrome. strength. Supplementation of factor XIII-deficient plasma with the purified zymogen. J Clin Invest 1983;71:1336–41. 12. Gladner JA, Nossal R. Effects of cross-linking on the rigidity and proteolytic susceptibility of human fibrin clots. Thromb Conclusions Res 1983;30:273–88. 13. Gorkun OV, Veklich YI, Medved’ LV, Henschen AH, Weisel JW. Role of the aC domains of fibrin clot formation. Bio- Patients with dysfibrinogenemia demonstrated a high chemistry 1994;33:6986–97. rate of clinical manifestations compared to previous 14. Medved’ LV, Gorkun OV, Manyakov VF, Belitser VA. The role studies, mostly with a tendency for easy bruising and of fibrinogen aC-domains in the fibrin assembly process. bleeding. Although the overall percentage of thrombosis FEBS Lett 1985;181:109–12. in our population was at the expected level, arterial 15. McDonagh I. Dysfibrinogenemias and other disorders of thrombosis was unexpectedly high, as the majority of fibrinogen structure or function. In: Colman RW, Hirsh J, Marder VI, Clowes AW, George JN, eds. Hemostasis and thrombotic events were of arterial origin. The high ten- thrombosis, basic principles and clinical practice. 4th ed. dency for miscarriage is striking. Management of patients Philadelphia, PA: Lippincott Williams & Wilkins, 2001:877. may be complicated by the variablility in clinical manifes- 16. Francis JL, Armstrong DJ. Acquired dysfibrinogenemia in tations and combinations thereof. Therapeutic recom- liver disease. J Clin Pathol 1982;35:667–72. mendations should be evaluated on the individual basis 17. Martinez J, Palascak JE, Kwasniak D. Abnormal sialic acid of the patient. Very careful treatment recommendations content of the dysfibrinogenemia associated with liver dis- ease. J Clin Invest 1978;61:535–8. have to be provided for patients with simultaneous pres- 18. Llobet D, Borrel M, Vila L, Valve C, Felices R, Fontcuberta entation of both bleeding and thrombosis. J. An acquired inhibitor that produced the delay of fibrino- In conclusion, consistent correlation between varia- peptide B release in an asymptomatic patient. Haematolo- tions in clinical manifestations and laboratory results gica 2007;92:e17–9. could not be established in our study. Further studies of 19. Dear A, Brennan SO, Sheat MJ, Faed JM, George PM. dysfibrinogenemia may lead to a better understanding of Acquired dysfibrinogenemias caused by monoclonal pro- hemostasis and thrombosis dependent on the underlying duction of immunoglobulin g light chain. Haematologica 2007;92: e111–7. fibrinogen gene mutation and of optimal treatment 20. Haverkate F, Samama M. Familial dysfibrinogenaemia and options for patients. . Report on a study of the SSC Subcommittee on Fibrinogen. Thromb Haemost 1995;73:151–23. 21. Krammer B, Anders O, Nagel HR, Burstein C, Steiner M. References Screening of dysfibrinogenemia using the fibrinogen func- tion versus antigen ratio. Thromb Res 1994;76:577–9. 22. Desvignes P, Bonnet P. Direct determination of plasma 1. Doolittle RF. Structural aspects of the fibrinogen to fibrin con- fibrinogen levels by heat precipitation: a comparison of the version. Adv Protein Chem 1973;27:1–109. technique against thrombin clottable fibrinogen with spec- 2. Blombaeck B, Hessel B, Hogg D, Therkildsen L. A two-step trophotometry and radial immunodiffusion. Clin Chim Acta fibrinogen-fibrin transition in blood coagulation. Nature 1978; 1981;110:9–17. 275:501–5. 23. Dintenfass L, Kammer S. Re-evaluation of heat precipitation Article in press - uncorrected proof

Miesbach: Diagnosis of patients with dysfibrinogenemia 405

method for plasma fibrinogen estimation: effect of abnormal 26. Matsuda M, Sugo T, Yoshida N, Terukina S, Yamazumi K, proteins and plasma viscosity. J Clin Pathol 1976;29:130–4. Niwa K, et al. Structure and function of fibrinogen: insights 24. Roberts HR. The dysfibrinogenaemias. Br J Haematol 2001; from dysfibrinogens. Thromb Haemost 1999;82:283–90. 114:249–57. 27. Collet JP, Soria J, Mirshahi M, Hirsch M, Dagonnet FB, Caen 25. Galanakis DK. Inherited dysfibrinogenemia: emerging J, et al. Dusart syndrome: a new concept of the relationship abnormal structure associations with pathologic and non- between fibrin clot architecture and fibrin clot degradability: pathologic dysfunctions. Semin Thromb Haemost 1993; hypofibrinolysis related to an abnormal clot structure. Blood 19:386–95. 1993;82:2462–9.