Antithrombin Deficiency

Updated: Aug 22, 2019
  • Author: Bryan A Mitton, MD, PhD; Chief Editor: Perumal Thiagarajan, MD  more...
  • Print

Practice Essentials

Antithrombin III (henceforth referred to as antithrombin or AT) is a 58-kDa molecule belonging to the serine protease inhibitor (serpin) superfamily that plays a central role in anticoagulation and in regulating appropriate wound healing in mammalian circulation systems. Antithrombin inhibits several enzymes of the coagulation system, including factor IIa, IXa, Xa, and XIIa. 

Quantitative antithrombin deficiency Type I results from heterozygous point mutations or major gene deletions leading to low antithrombin antigen and activity levels. The more common qualitative Type II is characterized by normal antithrombin levels and reduced function and is further categorised into IIa, b, or c, depending on which part of the antithrombin molecule is affected by the mutation. [1, 2]  Antithrombin deficiency may also be acquired. All forms of AT deficiency increase the risk for venous thrombosis and, far less commonly, arterial thrombosis. 

Antithrombin functional assay is the preferred initial diagnostic study, in order not to miss Type II deficiency. If abnormal, antithrombin antigen levels will distinguish between the two types. Functional assays assessing inhibitory activity on Xa have a higher sensitivity than those assessing thrombin and some patients with Type II antithrombin deficiency have only slightly reduced or even normal function, thus increasing the complexity of its diagnosis. [2]

When patients with a known inherited AT deficiency who are experiencing an acute thrombotic event fail to respond to intravenous heparin, treatment with a direct thrombin inhibitor (eg, argatroban, dabigatran) is recommended. For a planned major operation, correction of antithrombin levels using antithrombin concentrate products is recommended. In acute severe trauma, some studies also suggest a beneficial effect with antithrombin replacement. 

For patient education information, see the Deep Vein Thrombosis Health Center.



Paul Morawitz at the University of Tubingen first coined the term antithrombin in 1905 to describe plasma’s ability to neutralize thrombin activity. In 1965, Olav Egeberg described the first family with thrombotic disease due to inherited antithrombin deficiency, providing convincing evidence of the clinical importance of antithrombin. [3] For a historical overview of antithrombin research, see the excellent review by Ulrich Abildgaard. [4]  




As its name implies, antithrombin was first characterized as an inhibitor of thrombin. Antithrombin also affects other serine proteases of the coagulation cascade. [5, 6, 7, 8]  A diagrammatic representation of the serine proteases with which antithrombin interacts is shown in the image below. Studies have shown that antithrombin also has anti-inflammatory actions that are independent of its effect on regulating coagulation. [9, 10, 11, 12]


Antithrombin sites of action. Antithrombin sites of action.

Antithrombin Function in Anticoagulation and Inflammation

Antithrombin belongs to the serpin family of inhibitors, which include heparin cofactor II (HCII), alpha2-antiplasmin, plasminogen activator inhibitor-1 (PAI-1), C1-inhibitor, and alpha1-antitrypsin. [13] Antithrombin forms a 1:1 irreversible complex with its target active enzyme, and the complex is cleared by the liver with loss of target enzyme activity.

The serpin family of proteins have a highly conserved molecular structure, with 3 beta-sheets and 9 alpha-helices. [14] A region known as the reactive center loop (RCL) protrudes above the core of the serpin molecule and has a sequence of amino acids that is complementary to binding sites in the active sites of the target proteases. Cleavage at the reactive center by target proteases results in the activation of a unique mechanism of inhibition. [15] Antithrombin exists in 2 forms: 90% as the alpha form that is glycosylated at 4 sites (Asn-96, Asn-135, Asn-155 and Asn-192) and 10% as the beta form that is not glycosylated at position Asn-135. [16]

Antithrombin is synthesized primarily in the liver. It is secreted into the plasma in the form of a molecule with a molecular weight of 58,200 kDa. The normal plasma level is 150 mcg/mL and the plasma half-life is approximately 3 days. Thus, even short periods of abnormal liver function may reduce antithrombin production, leading to potential thrombosis.

Plasma antithrombin is comprised of 432 amino acids, 6 of which are cysteine residues that form 3 intramolecular disulfide bonds. The major physiologic role of the molecule, as the name implies, is the inhibition of thrombin (factor IIa). Additional target proteases include activated factors X, IX, XI, and XII. [17] Antithrombin also serves to reduce factor VII activity by accelerating the dissociation of the factor VIIa-tissue factor complex and preventing its reassociation. [8]

The mechanism of inactivation of serine proteinases occurs in two steps, with an initial weak interaction followed by a conformational change that ‘traps’ the protease. This mechanism is depicted in the image below.

Antithrombin (AT) neutralizes the enzyme (IIa) by Antithrombin (AT) neutralizes the enzyme (IIa) by forming a 1:1 stoichiometric complex (AT:IIa) between the arginine-serine sites of the 2 proteins. Binding of heparin to lysyl residues on AT results in a conformational change in AT, which makes it more available to bind thrombin (IIa), IXa, and Xa, thus markedly accelerating the rate of enzyme-inhibitor complex formation. AT also neutralizes XIa and XIIa.

Antithrombin (AT) neutralizes the enzyme (IIa) by forming a 1:1 stoichiometric complex (AT:IIa) between the arginine-serine sites of the 2 proteins. Binding of heparin to lysyl residues on AT results in a conformational change in AT, which makes it more available to bind thrombin (IIa), IXa, and Xa, thus markedly accelerating the rate of enzyme-inhibitor complex formation. AT also neutralizes XIa and XIIa. Transformation to the final complex involves formation of a highly stable bond between the Arg393 residue on antithrombin and the catalytic Ser residue(s) on thrombin. The formation of the antithrombin-proteinase complex is accelerated by heparin and related glycosaminoglycans, reviewed elsewhere.

In vitro studies have established the relative rates of thrombin generation and neutralization, but a study by Undas et al quantified the changes in the rate of activation and inactivation of several hemostatic factors in blood serially sampled from a bleeding time cut. [18] In this in vivo test system with an active, ongoing interaction between blood components and the injured vessel wall in flowing blood, it was noted that thrombin-antithrombin (TAT) complexes started increasing within 30 seconds of the bleeding time cut and reached a maximum by 180 seconds.

The pattern of increase was typical of the two phases of activation, which have been described in other models of thrombosis, with an initial 60- to 90-second initiation phase followed by a subsequent propagation phase, during which activation reaches its maximum level. [18] In the healthy volunteers, under basal conditions, the amount of thrombin formed exceeded TAT formation at all time points tested until bleeding stopped.

TAT complexes formed following the neutralization of thrombin by antithrombin have been used as a surrogate marker for thrombin generation; serial changes in TAT levels have been used to determine alterations of the extent of hemostatic activation in the course of a disease or to assess the impact of specific therapy (eg, the effect of heparin in potentially treating d isseminated intravascular coagulation).

Heparin cofactor II (HCII) is another physiologic protein inhibitor of hemostasis that appears to contribute about 20-30% of plasma heparin-cofactor activity in the presence of large amounts of heparin; HCII does not, however, contribute to anti–factor Xa activity. Therefore, it has been suggested that, in the assessment of the true heparin cofactor activity of antithrombin, the anti–factor Xa activity of antithrombin be measured within 30 seconds of incubation with factor Xa in the presence of small amounts of heparin in order to exclude the contribution of HCII to this assay.

The use of low doses of heparin in the test system and the use of factor Xa rather than thrombin allows for an accurate assessment of antithrombin's heparin cofactor activity with avoidance of the contribution of HCII to this assessment. Thrombomodulin, an endothelial cell receptor for thrombin, also binds antithrombin and accelerates its anticoagulant effect. In a purified system, tissue factor pathway inhibitor (TFPI) also appeared to potentiate the ability of antithrombin to neutralize activated coagulation factors.

Independent of its anticoagulant properties, antithrombin also exerts anti-inflammatory and anti-proliferative effects. A number of studies have documented the ability of antithrombin to inhibit leukocyte rolling and adhesion, which is thought to be at least partly due to the release of prostacyclins from endothelial cells. [19]

Oelschlager et al have shown that antithrombin produces a dose-dependent reduction in both lipopolysaccharide and tumor necrosis factor (TNF)–alpha activation of nuclear factor kB (NF-kB) in cultured monocytes and endothelial cells. [20] As a result, the synthesis of proinflammatory mediators such as interleukin (IL)-6, IL-8, and TNF is decreased, leading to an anti-inflammatory effect.

A number of studies have also shown that cleaved antithrombin has potent antiangiogenic and antitumor properties. Larsson and colleagues have shown that fibroblast growth factor (FGF)-induced angiogenesis in the chick embryo and angiogenesis in mouse fibrosarcoma tumors is inhibited by treatment with latent antithrombin. [21] There is literature to suggest that latent antithrombin may also induce apoptosis of endothelial cells by disrupting cell-matrix interactions.


Antithrombin Gene Structure

The gene for antithrombin, SERPINC1, is located on chromosome 1 band q23.1-23.9, has 7 exons and 6 introns, and is 13.5 kilobases (kb). The promoter region does not have a TATA or CAAT box. A control element at the 5' flanking region is thought to be critical for efficient synthesis of antithrombin, with homology to an enhancer of murine and human genes. The mRNA is 1567 nucleotides long and has an approximately 175 base pair (bp) 3' untranslated region. Two modes of splicing of the primary transcript are feasible at 2 sites in the first intron; the result is either a full native antithrombin molecule or a truncated product with a portion left within the cell.

Patients with AT deficiency, either inherited or acquired, are predisposed to serious venous and arterial thrombotic disease due to prolonged circulation and activity of activated coagulation factors. This increases the risk of thrombus formation at sites that fulfill Virchow's postulates (stasis, alteration of coagulability of the blood, and vessel wall damage). Even a 50% reduction in the level of antithrombin activity is sufficient to ‘tilt’ the coagulation system in favor of thrombosis.

The most common thrombotic manifestations in patients with antithrombin deficiency (AT deficiency) include lower extremity deep venous thrombosis, with recurrent VTE being common. [22] Other sites of thrombosis include the inferior vena cava, hepatic and portal veins, and renal, axillary, brachial, mesenteric, pelvic, cerebral, and retinal veins. Arterial thrombosis is far less common.

Despite their increased incidence of thrombosis, individuals with antithrombin deficiency have a normal life expectancy. The European Prospective Cohort on Thrombophilia (EPCOT) study recently looked at mortality in groups of various thrombophilia patients, including antithrombin deficiency, compared with a control group from March 1994 to December 2006. [23] Overall, they found no increased risk of death in individuals with thrombophilia. During the study period, 6.6% of patients with antithrombin deficiency died, compared with 5.1% of control patients. Additionally, they found the hazard ratio to be 1.65 with a confidence interval 0.91 to 2.93.

Inherited Antithrombin Deficiency

Inherited antithrombin deficiency (AT deficiency) can be broadly classified into two types.

Type I antithrombin deficiencies are heterozygous mutations that lead to a complete loss of the mutant antithrombin protein, resulting in immunologic and functional levels that are 50% or less than normal. The genetic basis of type I mutations includes major gene deletions or point mutations, with point mutations being more common. The mutations appear to cause a quantitative reduction in antithrombin synthesis by various processes, including premature termination of translation, aberrant RNA processing, and production of unstable antithrombin molecules that have shortened plasma half lives. [16]

One report described 22 novel mutations in the antithrombin gene, of which 9 missense mutations resulted in type I deficiency and led to low antithrombin activity and antigen levels. Clinically these mutations were all associated with venous thrombosis occurring before the age of 32 years. [24] Homozygous type I antithrombin deficiency (AT deficiency) is almost always fatal in utero. [17]

Type II antithrombin deficiencies are typically the result of single amino acid changes that result in functional deficits in a molecule that is otherwise normally synthesized and secreted into the plasma. The variant antithrombin molecules may have abnormalities at the reactive site (Type IIa) or the heparin binding site (Type IIb). Most cases of type II antithrombin deficiency are also heterozygous, although rare cases of homozygous type II deficiency have been described. [17]

A third category of type II (Type IIC) antithrombin deficiency also exists, in which multiple or "pleiotropic" abnormalities affect the reactive site, the heparin binding site, or the plasma concentration. Type II heparin binding site variants are not associated with a high risk of thrombosis unless the affected individual is a homozygote. [16]

A number of mutations in AT have been molecularly characterized. For example, the heterozygous form of a commonly inherited variant of antithrombin affecting the heparin-binding site (HBS) is not a risk factor for thrombosis. However, several cases of patients with homozygous mutations in the Heparin Binding Site (HBS) region of the antithrombin gene have been published, and homozygosity is associated with earlier presentation of thrombotic disease. [25] Two of these cases were shown to be associated with arterial thrombotic disease.

On the other hand, the replacement of the normal threonine-85 (Thr-85) by a nonpolar methionine (known as Antithrombin-Wibble) results in a mild adult-onset thrombotic disease, whereas replacement of the same Thr-85 by a polar lysine (known as Antithrombin-Wobble) results in early onset of thrombosis in childhood. Interestingly, fevers can trigger conformational stress on the Antithrombin-Wobble protein and favor thrombosis.

Finally, a homozygous type of antithrombin deficiency (antithrombin III Kumamoto) has been reported to be present in a family with consanguinity. It was shown to be associated with arterial thrombotic disease. The patient developed cerebral arterial thrombosis at age 17 years and subsequently developed venous thrombosis.

A current listing of mutations affecting the antithrombin gene is available at the Antithrombin Mutation Database. [26] A review of published mutations indicates that they are distributed throughout the molecule, with reactive center defects having the biggest impact on the potential for thrombosis, and heparin-binding defects carrying the least thrombotic risk.

Although it is well-recognized that inherited antithrombin deficiency (AT deficiency) confers a higher risk of coagulopathy than inherited deficiencies of protein C deficiency or p rotein S deficiency there is unpredictable variability in the incidence and severity of thrombotic manifestations in patients with inherited antithrombin deficiency. A population-based case control study found a 5-fold increased risk of thrombosis when antithrombin deficiency was associated with another genetic defect that predisposes to thrombosis. [27, 28] This risk increased to 20-fold when antithrombin deficiency was coupled with another acquired risk factor for thrombosis. [27] Co-inherited disorders include Activated Protein C Resistance (Factor V Leiden), protein C or S deficiency, thrombomodulin gene mutations, methylene tetrahydrofolate reductase (MTHFR) deficiency, and high lipoprotein (a) levels.

In families with inherited antithrombin deficiency, thrombotic complications often begin in the second decade of life. Approximately 40% of these events seem to be spontaneous in nature, with no clear provoking event such as major trauma, surgery or prolonged immobility. In the remaining 60%, additional precipitating factors, such as oral contraceptive use, pregnancy, labor and delivery, surgery, or trauma, may precipitate the thrombotic event. [29]


Acquired causes of antithrombin deficiency


For healthy full-term neonates, serum AT levels are typically >50% lower than adult reference values. Newborns do not have the thrombotic tendency noted in adults with similarly reduced values because of simultaneous reductions in their procoagulant levels and perhaps due to a protective role of alpha2-macroglobulin as a thrombin inhibitor in the neonate and in childhood. Premature infants have even lower serum levels. [30]

AT levels in the newborn rise to approximately 60% of that of adult levels 1 month after birth and reach normal value at approximately 3 months. [30] Genetic mutations can influence this level, but the superimposition of serious illnesses, can further reduce antithrombin due to increased consumption or decreased production.

Acute respiratory distress syndrome is a known cause of antithrombin deficiency and itself is a major cause of both morbidity and mortality in the newborn. Extracorporeal membrane oxygenation used in the treatment of respiratory failure can be associated with reduced antithrombin levels and increased thrombotic events. Other causes of acquired reductions of antithrombin in neonates include sepsis, asphyxia, liver disease, other causes of DIC, and maternal preeclampsia or eclampsia. [31, 32]


There is little strong clinical evidence that reduction of antithrombin occurs during normal pregnancy; one Scandinavian study reported that antithrombin levels were lower during the third trimester of pregnancy and in the postpartum period, but there has been no report specifically linking thrombosis to an acquired deficiency in AT. [33] Pregnancy-induced antithrombin deficiency, however, is more likely to be seen in twin and triplet pregnancies. [34]

Diseases associated with pregnancy, such as hypertension of pregnancy, eclampsia, liver dysfunction characterized by elevations in liver enzymes, and DIC, also reduce antithrombin levels. In these conditions, low-grade activation of coagulation with consumption of antithrombin is evident before gross deterioration of coagulation parameters occurs. [31, 35]

Pregnancy-induced AT deficiency has been shown to increase the risk of liver dysfunction independent of thrombocytopenia. [36]

Liver disease

Synthesis of antithrombin and other physiologically important inhibitors of hemostasis, synthesis of procoagulants, and clearance of activated coagulation factors are all regulated by the liver. Thus, the liver plays a central role in hemostasis.

The severity of liver disease correlates with reductions in antithrombin antigen levels. These reductions are not only due to impaired synthesis, but also to an element of increased consumption, particularly when additional risk factors, such as sepsis, surgery, and hypotension, are present in patients with chronic liver disease.

Patients with acute, massive hepatocellular injury and elevations of liver enzyme levels can often have a significantly larger component of a consumptive process than patients with slowly progressive end-stage liver disease. Because of the decreased synthesis of inhibitors as well as the decreased ability to clear activated coagulation factors, patients undergoing orthotopic liver transplantation predictably develop DIC with reduction inantithrombin levels.

Kidney disease

Importantly, patients with nephrotic syndrome lose antithrombin in the urine, resulting in reduced plasma levels, and they are at higher risk for thrombotic events. Conversely, patients with inherited antithrombin deficiency may develop renal failure due to renal vein thrombosis or due to glomerular deposition of fibrinogen. The degree of compromise in renal function may be such that these patients need renal replacement therapy. Furthermore, as renal dysfunction progresses, these patients lose increasing amounts antithrombin in the urine and, thus, become even more prone to develop thrombotic episodes. [37, 38]

Bone marrow transplantation

Veno-occlusive hepatic disease is seen in patients who undergo bone marrow transplantation, particularly in unrelated-donor transplantations, and it is associated with the development of microthrombi in the terminal hepatic venules. This results in rapid, marked deterioration of liver function, causing a coagulopathy characterized by the reduction in the level of antithrombin and, consequently, significant morbidity and mortality.


Interest in the role of antithrombin deficiency in the setting of sepsis and the critically ill patient has been growing. There appears to be a correlation between the severity of illness and the degree of antithrombin reduction. [17] However, to what extent the depletion of antithrombin affects the clinical condition of such patients, or whether a reduction in the levels of antithrombin is merely a marker of inflammation and illness, remains to be determined.

Mesters et al in 1996 demonstrated a correlation between marked reduction in serum antithrombin levels and poor outcomes in septic patients. [39] A number of studies thereafter suggested the use of antithrombin supplementation in patients with severe sepsis and septic shock. [40]

However, the KyberSept trial, which was published in 2001 and was the largest randomized controlled trial of severely septic patients treated with antithrombin supplementation, failed to demonstrate any significant beneficial effects on mortality at 28 days. [41] Of note, a subgroup of patients with severe sepsis and high risk of death with concurrent diagnosis of DIC were found to have a significant reduction in mortality when given antithrombin. [42]

In general, a number of studies regarding the use of antithrombin as a treatment in the intensive care setting have overall concluded that, although there may be some benefit to such therapy, highly supraphysiologic doses of antithrombin are necessary, and the concurrent use of any form of heparin negates the benefit that may be derived from antithrombin treatment in this setting. [17]

More recently, Tagami et al in a large, retrospective database analysis demonstrated decreased 28-day mortality in patients with severe pneumonia and sepsis-related DIC who were given therapeutic antithrombin. [43] Additionally, a small randomized controlled trial studying the use of antithrombin to treat DIC in patients with sepsis demonstrated increased recovery rates from DIC, but lacked adequate power to detect a reduction in 28-day mortality. [44]

Nonetheless, the use of supplemental antithrombin in septic patients remains controversial. Further analysis with large, randomized control studies will be required before definitive recommendations can be made.

Drug-induced reduction in antithrombin levels

Heparin, given by intravenous or subcutaneous routes, causes an approximately 30% reduction in antithrombin levels, presumably due to rapid clearance in vivo of heparin-antithrombin complexes. Plasma samples to determine baseline antithrombin levels must therefore not be drawn after exposure to heparin.

A large body of literature shows that estrogens/oral contraceptives can also reduce antithrombin levels, potentially resulting in hypercoagulability (See Hypercoagulability - Hereditary Thrombophilia and Lupus Anticoagulants Associated With Venous Thrombosis and Emboli).

Finally, AT deficiency has also been described with asparaginase therapy; this occurs by suppression of production of AT in the liver as part of the mechanism of action of this chemotherapeutic agent. [45, 46]



Antithrombin deficiency can be acquired and inherited; the inherited condition has a prevalence of 1 in 500–5000. [47]  In patients who develop venous thrombosis, the prevalence of hereditary antithrombin deficiency is between 1:20 and 1:200. [17]  Among the subtypes of antithrombin deficiency, type II antithrombin deficiency is at least twice as common as type I antithrombin deficiency in the general population. [48]  However, in symptomatic patients, type I antithrombin deficiency represents about 80% of the total cases, indicating that these individuals are more predisposed to VTE events than individuals with type II deficiencies. [49]

Antithrombin deficiency is not restricted to any particular ethnic group and has been found in many countries. In a study of 4000 Scottish blood donors, the prevalence of type I antithrombin deficiency was found to be 0.2/1000 and that of type II heparin binding site antithrombin deficiency was found to be 2-3/1000. [50]

Although no overt racial predilection for antithrombin deficiency is known, the literature, especially from the Far East, has described the presence of novel mutations in the antithrombin gene that have observed in thrombophilic patients in specific population groups. [51, 52]

Both men and women can present with the inherited disorder and clinical manifestations of antithrombin deficiency (AT deficiency) are evident at an early or later age, depending on the severity of the inherited genetic defect and also on the co-inheritance or presence of other thrombophilic mutations, drugs, or diseases.

The severely affected homozygous form of antithrombin deficiency may lead to spontaneous fetal loss, babies born small for their gestational age due to a small placenta secondary to thrombosed placental vessels, or severe thrombotic problems at birth. In other instances, thrombotic manifestations may start in the teenage years. 

Acquired antithrombin reductions are usually secondary to other illnesses or drugs.



Patients who are heterozygous for type I or II antithrombin (AT) deficiency develop significant thromboembolic complications, generally involving the deep veins. The lifetime risk of developing venous thromboembolism (VTE) depends on the subtype of antithrombin deficiency. In patients with type I inherited AT deficiency, the risk of thrombosis is estimated to be 1% per year, starting at age 15 years. The overall lifetime risk of developing a thrombotic event in patients with type I inherited AT deficiency is estimated to range from 50% to 85%.

In patients with type II antithrombin deficiency, the risk of developing VTE is higher in those patients who have reactive site defects as compared to heparin-binding site defects. Estimated lifetime risk of thrombosis in type II mutations has been reported to range from 6 to 20%, depending on the mutation site. [47, 16]

Patients may develop recurrent VTE at an early age and, if the condition is unrecognized or inadequately treated, they may die from such events. Long-term consequences, such as chronic leg ulcerations, severe venous varicosities, and postphlebitic syndrome, are common from repeated episodes of VTE, which cause significant morbidity. The prognosis of patients with reductions in antithrombin as part of other systemic disorders depends on the underlying disorder.

The frequency of arterial thrombotic complications is low in patients with antithrombin deficiency. However, mutations leading to arterial thromboses have been described.

The incidence of pregnancy-related VTE in women with antithrombin deficiency in early reports may have overestimated due to methodologic limitations such as selection bias in family studies. Subsequent studies have suggested a much lower level with ranges between 0.08-15.8% for risk  of initial VTE during. [53]  However, a family history of VTE significantly increases risk during pregnancy. VTE is the leading cause of direct maternal death and thrombotic complications during embryogenesis can lead to a variety of developmental abnormalities. [2]

The risk for VTE is also increased in women taking oral contraceptives (1.2-4.4%) or hormone replacement therapy (2.5-5.1%). [53]

Nephrotic syndrome has been associated with reductions in antithrombin and an increased incidence of venous thrombosis (renal vein, 60%; VTE, 40%) with only a 3% incidence of arterial thrombosis.


Serious long-term morbidity can result from the following issues:

  • Venous and arterial thromboembolic events

  • Postphlebitic syndrome due to extensive DVT as the first event

  • Recurrent VTE due to the discontinuation of oral anticoagulants

  • Sudden death due to the lack of prophylaxis in high-risk circumstances

  • Atypical site thrombosis such as Budd-Chiari syndrome

  • Bowel ischemia due to mesenteric vein thrombosis