Objective.—To review the current understanding of the pathophysiology of antithrombin deficiency and its role in congenital thrombophilia. Recommendations for diagnostic testing of antithrombin function and concentration, derived from the medical literature and consensus opinions of recognized experts in the field, are included. These recommendations specify whom, how, and when to test.
Data Sources.—Review of the published medical literature.
Data Extraction and Synthesis.—A summary of the medical literature and proposed testing recommendations were prepared and presented at the College of American Pathologists Conference XXXVI: Diagnostic Issues in Thrombophilia. After discussion at the conference, consensus recommendations presented in this article were accepted after a two-thirds majority vote by the participants.
Conclusions.—Antithrombin deficiency is an infrequent genetic abnormality that may be a significant contributing cause of thrombophilia. Antithrombin deficiency also may be observed in conjunction with other genetic or acquired risk factors. Assay of antithrombin plasma levels is appropriate in the laboratory evaluation of individuals with thrombophilia, preferably using a functional, amidolytic antithrombin assay. The diagnosis of antithrombin deficiency should be established only after other acquired causes of antithrombin deficiency, such as liver disease, consumptive coagulopathy, or heparin therapy, are excluded. A low antithrombin level should be confirmed with a subsequent assay on a fresh specimen, and family studies may be helpful to establish the diagnosis. Antigenic antithrombin assays may be of benefit in subclassification of the type of antithrombin deficiency and to confirm the decreased antithrombin level in patients with type I deficiency.
Antithrombin is an important regulator of the coagulation cascade in its function as a serine protease inhibitor, due to inhibition of thrombin and multiple other coagulation proteases. The description of antithrombin deficiency in 1965 established the first genetic association between deficiency of a natural anticoagulant and clinical venous thrombosis.1 This description set into motion the search for other genetic causes for venous thrombosis, with subsequent reports of deficiencies of protein C in 19812 and protein S in 1984.3,4 Today, the complexity of thrombophilia has grown with appreciation that multiple inherited and acquired risk factors may interact to result in a clinically thrombotic phenotype. Thus, not only are rare genetic deficiencies of the natural anticoagulants, such as antithrombin, important, but more common genetic polymorphisms and gain-of-function abnormalities in procoagulant proteins may interact to predispose a patient to thrombosis as well. This article reviews the pathophysiology of antithrombin as a serine protease inhibitor, discusses congenital and acquired deficiencies of antithrombin, and evaluates laboratory assays for their ability to determine antigenic and functional levels of antithrombin. The epidemiologic evidence for the role of antithrombin deficiency in venous thrombosis are summarized, and recommendations for laboratory testing of antithrombin are outlined.
Antithrombin, a heparin cofactor and member of the serine protease inhibitor (serpin) gene family, is an important protease inhibitor that regulates the function of several serine proteases in the coagulation cascade.5,6 Several different antithrombin activities in plasma were reported during the first half of the 20th century,7,8 leading to the classification of antithrombins I through IV.9 It was subsequently shown that these various antithrombin activities were actually the function of one molecule, antithrombin III, whose name was shortened to simply antithrombin at the 1993 meeting of the International Society on Thrombosis and Haemostasis.10,11
The mature antithrombin molecule has a molecular weight of 58 200 d with 432 amino acids,12,13 including 6 cysteine residues that form 3 intramolecular disulfide bonds (cysteines 8 with 128, 21 with 95, and 247 with 430) (Figure 1).13 There are also 4 asparagine residues (95, 135, 155, and 192) that are glycosylated with sialylated oligosaccharides in 90% of plasma antithrombin (α-antithrombin isoform).14 Approximately 10% of antithrombin (called the β-antithrombin isoform) has enhanced heparin affinity due to lack of glycosylation at Asn135.15,16 The plasma concentration of antithrombin is 112 to 140 mg/L,17 with a half-life of 2 to 3 days.10
The gene for antithrombin is located on human chromosome 1 (q23–25).18 The antithrombin gene spans 13 480 bp of DNA from the transcription start site to the poly A signal and has 7 exons (Figure 1).19–21 There are 9 full-length and 1 partial-length Alu repeat elements within the introns.20 The transcription start site is located 72 bp 5′ to the ATG initiation codon.22 There are several sequence polymorphisms of the antithrombin gene, including enzyme cutting site polymorphisms, a 76-bp length polymorphism at the 5′ end of the gene,23 and (ATT) repeat polymorphisms in the Alu 5 and Alu 8 repeat elements.19 The liver is the primary source of antithrombin synthesis and posttranslational glycosylation.22 Little is known about the transcriptional regulation of the antithrombin gene,24 but 2 regions flanking exon 1 are thought to regulate transcription through interaction with hepatocyte nuclear factor 4 and CCAAT enhancer-binding protein.25 Antithrombin is not an acute phase reactant, and its synthesis is not up-regulated during the inflammatory response.6,26
Antithrombin functions as a physiological inhibitor of the coagulation serine proteases thrombin27,28 and factor Xa.28,29 However, it is also able to inhibit factors IXa,29 XIa,30 XIIa,31 kallikrein,32 and plasmin.33 Antithrombin, like other serpins, is able to inactivate thrombin by forming a covalent 1:1 complex with the serine protease, a process termed suicide substrate inhibition.34 The tertiary structure of antithrombin includes a 5-stranded central β-sheet (the A-sheet), together with a heparin-binding helix and a mobile reactive site loop (Figure 2).35 The reactive site loop (P1–P17) of antithrombin includes a scissile P1–P1′ (Arg393–Ser394) bond that resembles the substrate for thrombin and other serine proteases.36,37 Once thrombin cleaves the bond, the protease is covalently linked to the P1 residue and the reactive loop peptide becomes mobile.38 The reactive loop peptide then hinges on residues P15–P10 and incorporates into the central β-sheet, becoming a sixth strand.39 This induces a conformational change in both antithrombin and thrombin, with a hingelike translocation of thrombin to the distal end of the antithrombin molecule and its inactivation due to geometric distortion of the active site of thrombin.40
In its native state, antithrombin inactivates the proteinases inefficiently, owing to conformational inaccessibility of the P1–P1′ bond. Inhibition is accelerated approximately 1000-fold by the binding of heparin to arginine residues in the D-helix of antithrombin, with a resultant conformational change of the P1–P17 loop and exposure of the P1–P1′ reactive center (Figure 2).41–43 The minimum heparin species necessary for inducing the conformational change in antithrombin has been determined to be a specific pentasaccharide sequence.44 While the pentasaccharide is sufficient to accelerate the inhibition of factor Xa, the inhibition of thrombin requires a bridging contribution from heparin and the formation of a trimolecular complex between antithrombin, thrombin, and the heparin species.5 The minimum length of heparin necessary to facilitate thrombin inhibition has been determined to be 18 saccharides in length, inclusive of the pentasaccharide.45 In the inhibition of both thrombin and factor Xa, the heparin species dissociates and is able to bind to a new antithrombin molecule once the covalent bond has formed between antithrombin and the protease.
Physiological activation of antithrombin is thought to occur on the luminal surface of endothelial cells, through binding of heparan sulfate molecules to ryudocan or syndecan.46,47 The presence of luminal heparan sulfate is thought to localize antithrombin in its active state to the endothelial cell surface, ready to scavenge local thrombin formation. This scavenger role of antithrombin may be of physiologic importance, as antithrombin is inefficient in inhibiting Xa and thrombin once they are incorporated into the prothrombinase complex,48 an effect that may be mediated by the prothrombin cleavage product, fragment 2 (F1+2).49
Pharmacologic activation of antithrombin by heparin for anticoagulation has been used for more than 60 years. Commercially available heparins are prepared from porcine intestinal mucosa or bovine lung, largely because they are a good source of heparin-rich mast cells. In general, most commercial heparin preparations are heterogeneous and have a molecular weight between 7000 and 25 000 d, with the pentasaccharide sequence comprising only approximately 30% of the mass. These larger heparin species are able to accelerate the inhibition of both factor Xa and thrombin, resulting in a narrow therapeutic window and risk of bleeding with excess heparin. Low-molecular-weight heparin species, prepared by chemical or enzymatic degradation, have a molecular weight of approximately 5000 d and primarily inhibit factor Xa.
A deficiency of antithrombin associated with venous thrombosis was first described in 19651 in a 4-generation family with early-onset thrombophilia. This association with apparent age-related risk of thrombosis was validated in other studies.50 Almost 70% of these patients had a thrombotic event before the age of 35 years, and 85% by age 50 years.50,51 Other studies have reported approximately 50% prevalence of thrombosis with heterozygous antithrombin deficiency.52 While these data indicated that antithrombin was a potent risk factor for thrombosis, other studies indicated that the heterozygous deficiency was not associated with a significant shortening of life span.53 Some individuals with antithrombin deficiency do not develop thrombosis; a possible explanation is low levels of fibrinogen or von Willebrand factor.54
Recognition of the high incidence of factor V Leiden and the G20210A prothrombin gene mutations in thrombophilia patients, especially those with early onset, indicated that thrombophilia is likely a multifactorial disorder and gave rise to the “multiple-hit” theory of thrombophilia.55 The potential interaction of other abnormalities in patients with antithrombin deficiency means that reported prevalence of thrombosis, for what at the time was characterized as a single genetic defect, must be considered potentially flawed.56 In a study of familial antithrombin deficiency, coinheritance of heterozygosity for factor V Leiden was shown to increase the prevalence of thrombosis and decrease the median age of first thrombosis from 26 years in antithrombin-deficient individuals to 16 years in individuals with both antithrombin deficiency and factor V Leiden.57 In the general population, factor V Leiden has a prevalence of approximately 5%, but studies of families with antithrombin deficiency have shown a higher prevalence of factor V Leiden heterozygosity, approximately 14%.57,58 Coinheritance of antithrombin deficiency and factor V Leiden may be more likely than coinheritance of other thrombophilic disorders because the genes for both antithrombin and factor V are located in chromosome 1, separated by approximately 3 to 11 centimorgans.55,57
Antithrombin deficiency is typically considered to have an autosomal dominant mode of inheritance, but some genetic heterogeneity does exist. It presents as a heterozygous state almost exclusively, the homozygous state being extremely rare and usually lethal, presenting with neonatal thrombosis.51 Individuals homozygous for antithrombin deficiencies affecting the heparin-binding site (type IIb) have been described; these individuals have onset of thrombosis in infancy or childhood and may have both arterial and venous thrombosis.51,59–62
The heterozygous state is still a rare defect and occurs in healthy populations with a frequency of about 1 (0.07%) in 140063 to 16 (0.16%) in 9669 blood donors.64,65 The frequency in the thrombophilia patient population is reported to vary between 1% and 8% (Table 1).66–76 This translates in most modern reports to an incidence of antithrombin deficiency as a cause of thrombophilia in about 1% to 2% of patients. The genetic defect occurs in many ethnic groups.77 The deficiency is associated with venous thrombosis almost exclusively and usually includes lower extremity deep venous thrombosis, pulmonary embolism, or thrombosis of unusual sites, such as the inferior vena cava, mesenteric, renal, or cerebral veins.51,78 Venous thrombosis often is spontaneous, but may be associated with pregnancy, infection, and surgery.50,78 An epidemiologic study has associated both low and high antithrombin levels with arterial disease, principally ischemic heart disease.79 Individuals with antithrombin deficiency have been shown to have an increased risk for thrombosis, ranging from a 5-fold to a 50-fold increase.80–82 Thrombosis in individuals with antithrombin deficiency is rare in the first decade, but the initial thrombotic event usually occurs before the age of 30.50,51 It is often associated with some predisposing feature, such as immobilization, dehydration, or the perioperative state. Along with the classic forms of venous thrombosis, another unique manifestation of antithrombin deficiency is heparin resistance.83
A number of classification systems for antithrombin deficiency have been proposed, but the Antithrombin Mutation Database system using a simple type I and type II defect has become the accepted way to classify the deficiency state.84–86
The type I state is characterized by a matched quantitative deficiency in the antigen and activity of antithrombin to about 50% of normal. About 80 point mutations have been recognized as causing a type I deficiency, and about 30 frameshift deletions or insertions have been reported, as well as 12 major gene deletions.86
The type II state is characterized by a normal antigenic level of antithrombin, with a low level of activity due to a dysfunctional protein. In patients with venous thrombosis and antithrombin deficiency, type II deficiencies account for approximately 40%.87 However, in an asymptomatic population of normal blood donors, type II deficiencies accounted for 88% of identified antithrombin-deficient individuals.65 This category has been further divided into 3 subtypes based on the nature of the genetic defect. Type IIa is caused by mutations affecting the reactive site. At least 12 mutations have been identified that interfere with the interaction between antithrombin and its target protease.84 These reactive site mutants have been further subdivided into 3 groups, depending on which active site residue (P2, P1, or P1′) is involved. At least 5 of these involve a CpG hotspot. Only some of the mutations have been well characterized.
Type IIb mutations involve the heparin-binding site. Again, at least 12 mutations have been found, with 5 of these involving CpG hot spots. These mutations in the heparin-binding site can be explained by 3 distinct mechanisms.88
1. Substitutions in the glycosaminoglycan binding site. Almost all of these substitutions involve arginine residues.
2. Substitution distortion of the heparin-binding site, in particular the Leu99Phe mutation on the helix C site, which is underneath the active heparin-binding site. This distortion apparently reduces binding by being in close enough proximity to the active heparin-binding site to cause structural change, which then leads to decreased binding to both heparin and pentasaccharide. At least 5 other distinct mutations appear to act in a similar way.
3. Substitutions involving glycosylation effects, which can alter antithrombin function and may increase the plasma level of β-antithrombin.89
Individuals with type IIb mutations involving the heparin binding site have a lower prevalence of thrombotic events (approximately 6%) compared to patients with type I or type IIa mutations.90 Additionally, as indicated previously, homozygous type IIb mutations are not lethal, but are associated with both venous and arterial thrombosis.59,62
Type IIc mutations include a pleiotropic group of at least 11 distinct mutants involving mutations in the sheet 1C.86,91 X-ray crystallography has suggested that the mutations in sheet 1C occur in areas near the reactive loop site.91 This may interfere with the mobility of the reactive loop site after heparin binding, which can influence the potential interaction with thrombin.91,92 Interestingly, these pleiotropic mutations show decreased levels of the mutated antithrombin in plasma, which may be caused by a combination of reduced synthesis and secretion, as well as increased catabolism.21,93
Antithrombin-deficient patients who develop venous thrombosis are treated in a fashion similar to any other patients who develop venous thrombosis. The usual treatment includes antithrombotic therapy with unfractionated heparin or low-molecular-weight heparin, followed by an oral vitamin K antagonist, such as warfarin. Some patients with very low antithrombin levels may be resistant to heparin therapy and may require increased doses of heparin or antithrombin concentrates. Patients receiving antithrombin concentrates should be monitored by antithrombin functional assays to assure that an adequate target antithrombin plasma concentration has been reached (usually 80%–120%).
Acquired Deficiency States
Many clinical conditions are associated with an acquired antithrombin deficiency state. These conditions can be separated in to 2 broad categories: impaired synthesis or loss of protein. Impaired synthesis is seen in conditions such as liver disease, malnutrition, premature infancy, inflammatory bowel diseases, or extensive burns.94,95 A loss of antithrombin due to consumptive coagulopathy is seen in disseminated intravascular coagulation (sepsis, shock), acute hemolytic transfusion reaction, thrombotic microangiopathy, malignancy, l-asparaginase therapy, and acute thrombotic episodes.96,97 Heparin therapy is often associated with decreased antithrombin levels due to the formation of covalent thrombin/antithrombin complexes and inactivation of antithrombin.98 However, it is unusual for antithrombin levels to decrease to levels lower than 50% to 60% during heparin therapy. Urinary protein loss during the nephrotic syndrome is also associated with low levels of antithrombin. Many other clinical conditions have associations with milder degrees of antithrombin deficiency, but probably doubtful associations with thrombosis. These other conditions include infections, vasculitis, hemodialysis, renal transplant rejection, sickle cell/thalassemia, and plasmapheresis, especially with albumin replacement. Estrogen and progesterone levels may influence the plasma level of antithrombin. Levels of antithrombin increase during menopause,99 but no additional changes are observed with hormone replacement therapy.100,101 Antithrombin levels increase with high-dose oral contraceptive therapy and decrease late in pregnancy, but the reference ranges for these populations are not significantly different from those for the general population.99
LABORATORY EVALUATION OF ANTITHROMBIN
The first assays developed for detection of antithrombin deficiency quantified the antigenic form of the molecule by radial immunodiffusion techniques or Laurell rocket electrophoresis.105–107 The immunologic methods are not widely used at the present time, but are useful for detection of patients with the rare type II defect. Little impetus has existed for the development of modern enzyme immunologic assay methods, and most laboratories still perform commercial radial immunodiffusion methods to quantify the antigenic form of antithrombin. While these methods are very specific, their sensitivity is limited. However, since most patients with congenital defects have about a 50% decrease in protein level,1,50 it is usually possible to obtain a correct result. Coefficients of variation (CVs) for radial immunodiffusion or other immunological methods are much larger than chromogenic substrate assays.108,109 In a 2000 College of American Pathologists proficiency survey, the antigenic antithrombin assays had a CV percentage of approximately 40% to 50%, while the amidolytic assays had much lower CV percentages, approximately 9% to 14%.108
Almost all current methods measure functional levels of the antithrombin protein by use of synthetic substrate technology using predominantly amidolytic methods. This technique employs a synthetic peptide that mimics the natural target substrate of the enzyme, which is attached to a chromogenic group at or near the cleavage site.110,111 In this assay, patient plasma is incubated with an excess of thrombin in the presence of heparin. In the first phase of the reaction, the antithrombin neutralizes the thrombin in the presence of heparin. The remaining thrombin, which is inversely proportional to the amount of antithrombin in the patient plasma, is then quantitated by the cleavage of para-nitroaniline from the peptide substrate at 405 nm.112 The assay is easily automated and can be performed on either a coagulation analyzer or almost any chemistry instrument. Some assays use inhibition of factor Xa rather than thrombin to decrease the contribution from other proteins, such as heparin cofactor II.113
Since most amidolytic assays have the potential for nonspecificity due to substrate cleavage by other proteases, the newer commercial antithrombin assays have protease inhibitors, such as aprotinin, which minimizes nonspecific substrate cleavage, and bovine thrombin, which is resistant to any impact of heparin cofactor II.114,115 The amidolytic antithrombin assays are very sensitive and specific, and are not influenced by the presence of heparin in the patient plasma.116 Most amidolytic assays are not able to distinguish patients with type IIb deficiency from the other categories, but this distinction is of doubtful clinical significance, except that individuals with type IIb deficiency may have a lower risk of thrombosis.90 Interassay CV percentages of 3% to 5% are possible, although CV percentages of approximately 10% were observed in a recent College of American Pathologists survey.108 The assay has a positive predictive value of 96% in patients with congenital antithrombin deficiency states.
A 2-SD normal range for antithrombin activity obtained from 9669 normal blood donors was determined to be 83% to 128%.99 The normal range was not significantly different for males or females, in females on oral contraceptive therapy, during pregnancy, or with smoking.99,117 Antithrombin levels are lower in neonates and increase to adult ranges by approximately 1 year; levels then are slightly increased compared to adult levels up until approximately age 16 years.94,118
Recent assay developments include the description of a global clotting-based assay for the factor II/antithrombin system and the development of a clot-based assay for antithrombin quantitation, which uses a heparinized antithrombin-deficient plasma and is performed exactly like a traditional coagulation factor assay.119 These types of assays may be useful in laboratories that do not have instrumentation to perform chromogenic assays. Prior clotting antithrombin assays were difficult to perform and subject to variability as they used thrombin, so the plasma had to be defibrinated before testing or a serum sample was used.120,121
Genetic analysis has been important in identifying the various specific mutations mentioned, but for the diagnostic clinical laboratory and routine medical practice, these data are likely to have little relevance beyond the classification of types I and II.
Proper performance of functional antithrombin assays should also include consideration of preanalytic variables, establishment of an accurate reference range, and validation of the calibrator used to establish the standard curve. Due to ex vivo coagulation activation, coagulation testing specimens may be plagued by preanalytic variables, so careful collection and processing is required. Samples that are hemolyzed, lipemic, clotted, from individuals with a hematocrit greater than 55%, or that are underfilled are not acceptable for performing antithrombin assays or coagulation testing in general.122 Laboratory-specific reference ranges should be established by analysis of plasma samples from at least 40 normal individuals by calculating the population mean ± 2 SD from the mean.123 Calculation of laboratory-specific reference ranges is important, owing to variability in test methodology from site to site and variability in population parameters. Most amidolytic antithrombin results are calculated from a standard curve constructed using serial dilutions of a normal plasma calibrator, with results reported as percentage of normal. Manufacturers usually calibrate these plasmas against the current Antithrombin Standard from the National Institute for Biological Standards and Control,124 but laboratories should validate that the antithrombin activity of the calibrator provided by the manufacturer is accurate.
The issue of who should be tested is relatively easy in the case of a patient with a strong family history of thrombosis or in a young individual with thrombosis and no apparent family history.125,126 Also, recognition of the high frequencies of the factor V Leiden and prothrombin G20210A mutations in conjunction with the multiple-hit etiology of thrombophilia means that most patients now will have a panel of the more common thrombophilia markers analyzed during an evaluation.
Controversy still surrounds older thrombophilia patients (>50 years) with no family history or individuals with thrombosis associated with a predisposing event. Given that almost all antithrombin-deficient patients will have a thrombosis before this age, as well as the relative rarity of the condition, some clinicians would argue that it is not justified either on scientific or cost grounds to test such patients for antithrombin deficiency or any other congenital thrombophilic disorders.126,127 Much of this controversy relates to the fact that the thrombosis itself will be treated in the same way, regardless of whether the patient has an identified etiology.127,128
Almost all data indicate that testing patients for thrombophilia markers in the acute phase of a thrombotic event is not appropriate, given the variable consumption of factors like antithrombin and protein C during such times.129–131 However, in modern hospital practice with immense pressure for short or no hospitalization for venous thrombosis patients, for example, it may be the only time the patient can be tested before several months of anticoagulant therapy. If antithrombin testing is carried out during this time and yields normal results, it is unlikely the patient is antithrombin deficient. However, if the result is only modestly abnormal, then no diagnostic conclusion can be drawn, since intravenous unfractionated heparin therapy itself can cause lowering of the antithrombin level in plasma by about 25%.98,132
In an ideal situation, the patient should be evaluated at least 3 months after their event to determine potential antithrombin deficiency. The algorithm in Figure 3 may be helpful in establishing a diagnosis of antithrombin deficiency. If the initial antithrombin level is normal or elevated, antithrombin deficiency is unlikely. However, antithrombin levels may be increased in some patients on oral anticoagulation therapy, which could potentially mask an antithrombin deficiency. If the initial antithrombin level is low, then a confirmatory antithrombin test should be done on the patient using a repeat specimen. It may also be helpful to test first-degree relatives for antithrombin deficiency. These studies should confirm the antithrombin deficiency before the patient is considered to be truly antithrombin deficient. No patient should be diagnosed on 1 assay alone, no matter how profound the antithrombin deficiency.133 Acquired causes of antithrombin deficiency should be excluded. On the repeat specimen, both activity and immunological assays may be carried out to determine if the patient has a type I or type II antithrombin deficiency. However, since the anticoagulant therapy is the same for both type I and type II antithrombin deficiencies, the subclassification is not clinically necessary, except to address familial or epidemiologic concerns. Further characterization of the actual gene mutation is rarely performed because of lack of clinical necessity and the large number of potential mutations. Genetic identification of type IIb may be of epidemiologic interest, as it is associated with a lower thrombosis risk because the mutations do not involve the active site, but rather involve the heparin-binding site.
In patients with acquired antithrombin deficiency, the use and frequency of antithrombin testing will depend on the clinical situation. Patients with active disseminated intravascular coagulation may require testing every hour if they are actively bleeding, whereas a patient with a more indolent disease state, such as nephrotic syndrome, may require testing much more infrequently.
CRITERIA FOR DIAGNOSIS
Almost all patients with congenital antithrombin deficiency will have about a 50% reduction in the plasma antithrombin activity level to a range of 35% to 70%. There will be a parallel decrease in the antithrombin antigen level to about 50% of normal in the type I patients, with a normal antithrombin antigen level in most type II patients.
Given the precision of the current amidolytic activity assays, only in thrombophilia patients in whom the antithrombin level is mildly decreased (70%–80% or within 3 SD below the laboratory mean normal) should the diagnosis be questioned, as long as acquired causes of decreased antithrombin can be excluded. Almost no patients with antithrombin levels in this range will have thrombosis, and the laboratory should be concerned about preanalytic variables, such as short-draw samples, as a cause of such results. No patient should be “labeled” as having antithrombin deficiency at this level until repeat testing or some other form of confirmation is obtained. This stipulation is particularly important in laboratories that provide reference testing services, where abnormal antithrombin results can be returned to clients with minimal guidance as to the clinical significance of the abnormal result.
Antithrombin deficiency is an infrequent genetic abnormality that may be a significant contributing cause of thrombophilia, but that also may be observed in conjunction with other genetic or acquired risk factors (see “Summary” in Table 2). Specific recommendations for the inclusion of antithrombin testing in the general thrombophilia evaluation are included in the article by Van Cott et al,134 but when required, assay of antithrombin plasma levels preferably should be performed using a functional, amidolytic assay (see Table 3 for a summary of recommendations). Isolated testing for antithrombin is recommended when an individual from a family with known antithrombin deficiency requires testing. Isolated testing for antithrombin is also recommended as a confirmatory test when an abnormal antithrombin value was found in the initial test, either in the same or a different laboratory. Owing to its low prevalence in the general population, routine measurement of antithrombin is not recommended prior to starting oral contraceptive or hormone replacement therapy unless there is a family history of antithrombin deficiency. The diagnosis of antithrombin deficiency should be established only after other acquired causes of antithrombin deficiency, such as liver disease, consumptive coagulopathy, or heparin therapy, are excluded.131,135 A low antithrombin level should be confirmed with a subsequent assay on a new specimen, and family studies may be helpful in establishing the diagnosis. There is no need to routinely perform antithrombin antigen assays. However, antigenic antithrombin assays may be useful for distinguishing type I from type II antithrombin deficiency, and to confirm the decreased antithrombin antigen level in patients with type I deficiency (see Table 2). Pharmacologic agents (especially heparin) should be taken into consideration in interpretation of antithrombin results, but testing for antithrombin deficiency is best done at least 5 days after cessation of heparin therapy. It is preferable not to test for antithrombin deficiency during an acute event (thrombotic, surgical, etc). However, a normal value in the setting of an acute event excludes antithrombin deficiency.
Presented at the College of American Pathologists Consensus Conference XXXVI: Diagnostic Issues in Thrombophilia, Atlanta, Ga, November 9–11, 2001.
Reprints: Kandice Kottke-Marchant, MD, PhD, Department of Clinical Pathology, L30, Cleveland Clinic Foundation, 9500 Euclid Ave, Cleveland, OH 44195 (firstname.lastname@example.org)