The International Normalized Ratio
A Tool for Monitoring Warfarin Therapy
By Neil S. Harris, MBChB, MD, and William E. Winter, MD
Warfarin is a widely prescribed oral anticoagulant that acts by inhibiting vitamin K-dependent coagulation factors in blood. Too much warfarin, however, causes bleeding and can even result in death. Therefore, dosage of the drug must be individualized for each patient.
To avoid adverse events associated with warfarin such as excessive bleeding, clinicians regularly monitor patients’ anticoagulation status using prothrombin time (PT) and the international normalized ratio (INR). Individuals who are at risk for bleeding while receiving warfarin include the elderly, as well as patients with liver disease, congestive heart failure, or those on hemodialysis. Recent major surgery, malnutrition, hyperthyroidism, and many drugs, including amiodarone (an anti-arrhythmic), also increase the chances of bleeding. Furthermore, there are also important genetic factors that modulate the response to warfarin. Patients with such risk factors may benefit from more frequent monitoring, careful dose adjustment to desired INR, and shorter duration of therapy. Clinicians also commonly use PT and INR to assess patients’ hemostastic systems.
As laboratorians, we either perform these assays in the clinical lab or oversee testing performed at the point-of-care. This article summarizes the biological basis behind PT, highlights the derivation of the INR, and describes appropriate usage of these laboratory tests.
Coagulation In Vivo—An Overview
Coagulation factors are proteins that participate in the coagulation cascade responsible for clot formation. The reactions catalyzed by the enzymes of the coagulation cascade occur on cell surfaces, initially on tissue factor-bearing cells, and later on the surface of activated platelets (1). Activated clotting factor VII (FVIIa) and a cell-surface protein called tissue factor (TF) expressed on fibroblasts, smooth muscle cells, epithelial cells, and astroglia initiate coagulation in vivo. TF is not present on the surface of vascular endothelial cells, but damage to blood vessels and subsequent leakage of plasma into the extravascular compartment exposes TF to pre-activated FVIIa, which is constitutively present in small amounts in plasma (1, 2). When the integrity of the blood vessel is disrupted, factor FVIIa and TF form an active enzyme complex (FVIIa/TF) on TF-bearing cells that allows FVIIa to cleave and activate FX to FXa. FXa then interacts with FVa to cleave prothrombin, thereby generating thrombin. Thrombin in turn activates FVII and generates much larger amounts of FVIIa. Thrombin also cleaves two peptides, fibrinopeptides A and B, from fibrinogen to generate fibrin. This proteolytic cleavage creates a strong non-covalent interaction between fibrin monomers, leading to generation of a fibrin strand.
The FVIIa/TF complex also activates FIX to FIXa. The latter interacts with FVIIIa to form a very active FX-cleaving enzyme that generates abundant amounts of FXa (2). FVIIa/TF can activate FX directly or indirectly via generation of the FIXa/FVIIIa complex. Furthermore, thrombin generated in the early stages of the initiation phase will activate both FV and FVIII to form FVa and FVIIIa through proteolytic cleavage.
This in vivo model of coagulation has two phases: initiation, consisting of FVIIa/TF→FXa→FXa/FVa→prothrom-bin→thrombin, and propagation, consisting of FIXa/FVIIIa→FXa→FXa/FVa→prothrombin→thrombin (2) (Figure 1). Physical coagulum formation occurs as the propagation phase becomes dominant, approximately 4–5 minutes after vessel injury. Coagulum formation is not the end of the process since the propagation phase continues beyond this point and thrombin concentrations continue to climb. Formation of a visible coagulum occurs after only 3–5% of the total amount of thrombin is produced (2); therefore, the total amount of thrombin generated greatly exceeds the amount required to initiate fibrin polymerization for clotting.
In Vivo Coagulation Pathway
There are two important points at this stage of the discussion. First, the in vivo coagulation cascade occurs in whole blood and in the presence of blood platelets. The platelets provide an active phospholipid surface to which the coagulation factors bind. Thrombin also activates platelets, leading to platelet aggregation and formation of a platelet plug. Lab measurements of plasma-based coagulation factors, however, are performed after platelets are separated from the plasma. Therefore, lab analysis of patients’ coagulation status is not an exact replica of the in vivo interactions of the factors in the coagulation cascade.
Second, the true endpoint of the coagulation factor cascade appears to be generation of thrombin, not formation of fibrin (2). Thrombin deficiency is lethal and incompatible with life, while fibrinogen deficiency makes the individual much more prone to bleeding, but is seldom lethal.
Post Translational Modification of Clotting Factors
Clotting factors, like all proteins, consist of a number of distinct functional domains. One of the distinctive domains present in many, but not all of the factors, is Gla, which undergoes an enzyme-catalyzed carboxylation in the presence of oxygen and carbon dioxide. The enzyme γ-glutamylcarboxylase adds a carboxyl group to the glutamic acid side chains, thereby facilitating interaction of the factor with activated phospholipid surfaces and allowing the coagulation protein to bind to cell surfaces in a calcium-dependent fashion. The proteins that contain the Gla domain include FVII, FIX, FX, and prothrombin. The Gla domain is also found in proteins C and S that inhibit hemostasis.
This post-translation modification step is the target of warfarin’s inhibition of coagulation in vivo. In the absence of γ-glutamylcarboxylation, FVII, FIX, FX and FII remain functionally inactive.
The Role of Vitamin K in Coagulation
Vitamin K, a fat-soluble vitamin, is essential for the process described above. In the carboxylation step, oxygen is reduced to water and the electron donor is reduced vitamin K (vitamin K hydroquinone), which subsequently undergoes oxidation to vitamin K 2,3 epoxide (Figure 2). If this process continued unabated, all the vitamin K would be converted to the epoxide, and the reaction would stop. Vitamin K epoxide reductase subunit 1 encoded by the VKORC1 gene regenerates the reduced vitamin K hydroquinone.
Vitamin K Epoxide Reductase
Warfarin prevents blood from clotting by inhibiting vitamin K epoxide reductase, which interferes with carboxylation of glutamic acid residues on coagulation factors FII, FVII, FIX, and FX. Without this post-translational modification, the factors are inactive.
Warfarin Inhibition of Hemostasis
Warfarin prevents blood from clotting by inhibiting vitamin K epoxide reductase, the enzyme responsible for reactivating vitamin K. This inhibition prevents the carboxylation, and ultimately, activation of the Gla domain-containing coagulation factors. Although FII, FVII, FIX, and FX are still synthesized in the liver, the critical post-translational carboxylation is blocked. Not all coagulation factors are affected, however. FVIII and FV do not have a Gla domain, so their activity is not altered.
Patients on warfarin therapy will have both prolonged PT and activated partial thromboplastin time (aPTT) times, but PT is more sensitive to warfarin than aPTT. This is because the majority of clotting factors that participate in PT contain vitamin K-dependent Gla domains (FVII, FX and FII). In aPTT, the contact system (FXIIa, FXIa, kallikrein, and high molecular weight kininogen, all of which are vitamin K-independent), determines much of the elapsed time of the reaction. Warfarin’s effect on PT takes several days. Activated vitamin K-dependent factors present in the plasma must also go through their half-life cycle, which for some is 70 hours or more (Table 1), before the drug’s full effect is realized.
Half-life of Coagulation Factors
|Factor VII (FVII)
|Factor IX (FIX)
|Factor X (FX)
Some patients on warfarin therapy experience a transient hypercoagulable state early in the course of taking the drug. This is because both proteins C and S are vitamin K-dependent. Since protein C degrades FV and FVIII, it functions as an inhibitor of hemostasis. Protein C has a short half-life of ≤6 hours. Early in the course of warfarin anticoagulation, concentrations of Gla-carboxylated protein C may fall below a critical threshold, before the procoagulant factors reach a similar state. This is most likely to occur in individuals with low plasma concentrations of protein C. One of the possible consequences of this state is warfarin-induced skin necrosis, a painful and unfortunate outcome likely due to microvenular thrombosis.
Clinicians use prolongation of PT as a surrogate measure of the effectiveness of patients’ warfarin therapy. When PT is extended appropriately, the patient is at reduced risk of thrombosis and significant bleeding. Lab assessment of PT measures the biological activity of clotting factors FVII, FX, FV, FII and fibrinogen (FI).
Measurement of PT starts with the addition of thromboplastin and calcium ions to citrated plasma at 37°C. (See below for specimen description.) The first thromboplastins were derived from animal brain extracts and contained abundant amounts of TF and phospholipids. Today, laboratories use thromboplastins from partially-purified brain extract or synthetic thromboplastins containing a recombinant TF.
Thromboplastin and calcium initiate coagulation via a mechanism that is similar to the initiation of in vivo coagulation. This in vitro sequence of events is called the extrinsic pathway. First, the FVIIa/TF complex activates FX. FXa and FVa together cleave prothrombin to form thrombin, and the latter converts fibrinogen to fibrin strands. Although FVIIa/TF can activate FIX, the role of the FIXa/FVIIIa complex is overwhelmed by the very high concentrations of TF in the reaction, and cross-over from the extrinsic to the intrinsic pathway (the in vitro cascade involving FXII, FXI, FIX, and FVIII) essentially does not occur. Consequently, PT is insensitive to deficiencies of FVIII, FIX, FXI, and FXII.
There are three essential differences between PT and in vivo hemostasis. First, PT reactions occur in platelet-free plasma. Platelet adhesion and aggregation are removed from the series of events. Second, the endpoint of PT is the formation of a fibrin-based clot that can be detected optically or mechanically. The reaction contains very high concentrations of TF so that the endpoint of the reaction (fibrin generation) is typically achieved within 15 seconds. Finally, PT is not sensitive to defects in platelet function or deficiencies of FVIII and FIX found in individuals with hemophilia A and B, respectively.
Proper Specimens for PT and INR
For laboratory analysis of clotting time, phlebotomists should be instructed to draw blood into a tube containing liquid 3.2% or 109 mM trisodium citrate (Na3C6H5O7• 2H2O) at a ratio of nine parts blood and one part anticoagulant. Typically, labs perform all plasma-based coagulation tests on platelet-poor plasma, defined as a platelet count of < 10,000/μL (10 X 109/L). PT performed on fresh plasma samples is not affected by platelet counts up to 200,000/μL (200 X 109/L) (5); however, if the specimen is to be frozen for later testing it should be platelet-poor. Platelets release phospholipids on freeze-thawing, which will affect the clotting time. Guidelines from the Clinical Laboratory and Standards Institute recommend that PT specimens, uncentrifuged or centrifuged with plasma on top of cells, should be kept at 18–24°C for no longer than 24 hours (4, 5). If the clinician orders a concurrent aPTT, this time is reduced to only 4 hours.
The Role of INR
Because labs use thromboplastins produced by several methods and originating from different sources, PTs performed on the same specimen in different labs vary significantly. To standardize the results, the World Health Organization (WHO) devised a formula that uses the ratio of the patient’s PT results and the mean of the normal range: INR = [PTpatient/MNPT]ISI. The first step of the INR calculation is to “normalize” the PT by comparing it to the mean normal prothrombin time (MNPT), the geometric mean of the prothrombin times of the healthy adult population. In the second step, the ratio is raised to a power designated the ISI for international sensitivity index. The ISI is a function of the thromboplastin reagent compared with an international WHO standard and is derived from two groups of data: normal, healthy individuals and patients stabilized on vitamin K antagonists (Figure 3). Paired PT data are obtained from these samples using both the working PT reagent and the international PT standard. The slope of the line is the ISI.
The International Sensitivity Index of Thromboplastin
The ISI is a function of the thromboplastin reagent compared with an international WHO standard (after ref. 6).
If the thromboplastin reagent is identical to the WHO standard, the ISI will be 1.00. In practice, ISIs typically range from 0.9–1.8. ISIs are reasonably consistent for reagents produced by a single vendor, although there may be a significant lot-to-lot variation. It is vital that the laboratory check and confirm changes to the ISI with each thromboplastin lot. Labs should report INR values to one decimal place.
Although the aim of the INR is to standardize PT results for patients on warfarin therapy, it is not perfect. INRs performed on different instruments using different thromboplastin reagents will not be exactly the same (7, 8). For example, in a recent College of American Pathologists (CAP) proficiency test participant summary (7), the interassay % CV for all methods/reagents is 8% at an INR of 2.8 and 12% at an INR of 4.3. Acceptable performance for a CAP INR proficiency study is the mean of the peer group ± 20%.
A Need for Better Standardization
PT and INR are critical tests for patients on warfarin therapy. Although these tests rank among the most commonly performed laboratory tests, their biological basis and limitations are often poorly understood. While use of the INR has resulted in better agreement between labs, in reality, INR results from the same specimen performed on different analyzers with different thromboplastin reagents still are not the same.
Furthermore, although INR reporting has helped improve management of patients on anticoagulation therapy, it is not appropriate for certain groups of patients, including those who are just beginning anticoagulation therapy. Patients with liver disease should also not be monitored with INR (6). In liver disease, FV and fibrinogen may be present in blood at lower than normal concentrations in addition to the decreased levels of vitamin K-dependent factors and other abnormalities that affect clotting. Finally, patients who are being screened for clotting factor deficiencies should be evaluated using PT and aPTT results, but not with INR.
As laboratorians, we need to work to educate clinicians about the limitations of PT and INR in order to optimize proper use and interpretation of the results. Increased efforts to standardize reagent (10) and instrument performance will help laboratories report consistent INR values on patient samples and improve the value of the test.
- Harris NS, Winter WE, Ledford-Kraemer MR. Hemostasis: A Review And Methods Of Assessment In The Clinical Laboratory. 2010 IN PRESS. Contemporary Practice in Clinical Chemistry, Clarke, W. Ed. (AACC Press, Washington DC).
- Mann KG, Butenas S, Brummel K. The Dynamics of Thrombin Formation. Arterioscler Thromb Vasc Biol 2003;23:17–25.
- Hoffman M, Monroe DM. Coagulation 2006: A Modern View of Hemostasis Hematol Oncol Clin N Am 2007; 21:1–11.
- CLSI. Collection, Transport, and Processing of Blood Specimens for Testing Plasma-Based Coagulation Assays; Approved Guideline—Fifth Edition. CLSI Document H21-A5 (ISBN 1-56238-657-3).
- Hematology and Coagulation Checklist. 2010. College of American Pathologists, Northfield, Ill.
- Tripodi A, Caldwell SH, Hoffman M, Trotter JF, et al. Review article: the prothrombin time test as a measure of bleeding risk and prognosis in liver disease. Aliment Pharmacol Ther 2007;26:141–8.
- CAP INR Proficiency Study Participant Summary, Coagulation Limited, CGL-B 2010. 2010 College of American Pathologists, Northfield, Ill.
- Horsti J, Uppa H, Vilpo JA. Poor Agreement among Prothrombin Time International Normalized Ratio Methods: Comparison of Seven Commercial Reagents. Clin Chem 2005;51:553–560.
- Favaloro EJ, Adcock DM. Standardization of the INR: How Good Is Your Laboratory’s INR and Can It Be Improved? Semin Thromb Hemost 2008;34:593–603.
- van den Besselaar AMHP, et al. Thromboplastin standards, Biologicals (2010), doi:10.1016/j. biologicals 2010.02.012 [Article in press].
Neil S. Harris, MBChB, MD, is director of the Core Laboratory and a clinical associate professor in the Department of Pathology, Immunology, and Laboratory Medicine at the University of Florida College of Medicine, Gainesville. Email: firstname.lastname@example.org
William E. Winter, MD, is professor of pathology and pediatrics at the University of Florida in Gainesville. Email: email@example.com