The luminal surface of the endothelial cell ¹ is covered by the glycocalyx, a proteoglycan coat. It contains heparan sulphate and other glycosaminoglycans, which are capable of activating antithrombin, an important inhibitor of coagulation enzymes. Tissue factor pathway inhibitor (TFPI) is present on endothelial cell surfaces bound to these heparans but also tethered to a glycophosphoinositol (GPI) anchor. The relative importance of these two TFPI pools is not known. Endothelial cells express a number of coagulation active proteins that play an important regulatory role such as thrombomodulin and the endothelial protein C (PC) receptor. Thrombin generated at the site of injury is rapidly bound to a specific product of the endothelial cell, thrombomodulin. When bound to this protein, thrombin can activate PC (which degrades factors Va and VIIIa) and a carboxypeptidase which inhibits fibrinolysis (discussed later). Thrombin also stimulates the endothelial cell to produce tissue plasminogen activator (tPA). The endothelium can also synthesize protein S, the cofactor for PC. Finally, endothelium produces von Willebrand factor (VWF), which is essential for platelet adhesion to the subendothelium and stabilizes factor VIII within the circulation. VWF is both stored in specific granules called Weibel Palade bodies and secreted constitutively, partly into the circulation and partly toward the subendothelium where it binds directly to collagen and other matrix proteins. The expression of these and other important molecules such as adhesion molecules and their receptors are modulated by inflammatory cytokines. The lipid bilayer membrane also contains adenosine diphosphatase (ADPase), an enzyme that degrades adenosine diphosphate (ADP), which is a potent platelet agonist (see p. 434). Many of the surface proteins are found localized in the specialized lipid rafts and invaginations called ‘caveolae’, which may provide an important level of regulation.²

The endothelial cell participates in vasoregulation by producing and metabolizing numerous vasoactive substances. On the one hand, it metabolizes and inactivates vasoactive peptides such as bradykinin; on the other hand, it can also generate angiotensin II, a local vasoconstrictor, from circulating angiotensin I. Under appropriate stimulation the endothelial cell can produce vasodilators such as nitric oxide (NO) and prostacyclin or vasoconstrictors such as endothelin and thromboxane. These substances have their principal vasoregulatory effect via the smooth muscle but also have some effect on platelets.

The subendothelium consists of connective tissues composed of collagen (principally types I, III and VI), elastic tissues, proteoglycans and non-collagenous glycoproteins, including fibronectin and VWF. After vessel wall damage has occurred, these components are exposed and are then responsible for platelet adherence. This appears to be mediated by VWF binding to collagen. VWF then undergoes a conformational change and platelets are captured via their surface membrane glycoprotein Ib binding to VWF. Platelet activation follows and a conformational change in glycoprotein IIbIIIa allows further, more secure, binding to VWF via this receptor as well as to fibrinogen. At low shear rates (<1000 s⁻¹) platelet binding directly to collagen appears to dominate.³

Vasoconstriction

Vessels with muscular coats contract following injury, thus helping to arrest blood loss. Although not all coagulation reactions are enhanced by reduced flow, this probably assists in the formation of a stable fibrin plug by allowing activated factors to accumulate to critical concentrations. Vasoconstriction ^1,⁴ also occurs in the microcirculation in vessels without smooth muscle cells. Endothelial cells themselves can produce vasoconstrictors such as angiotensin II. In addition, activated platelets produce thromboxane A₂ (TXA₂), which is a potent vasoconstrictor.

Platelets

Platelets ^5,⁶ are small fragments of cytoplasm derived from megakaryocytes. On average, they are 1.5–3.5 μm in diameter but may be larger in some disease states. They do not contain a nucleus and are bounded by a typical lipid bilayer membrane. Beneath the outer membrane lies the marginal band of microtubules, which maintain the shape of the platelet and depolymerize when aggregation begins. The central cytoplasm is dominated by the three types of platelet granules: the δ granules, α granules and lysosomal granules. The contents of these various granules are detailed in Table 18.1. Finally there exist the dense tubular system and the canalicular membrane system; the latter communicates with the exterior. It is not clear how all these elements act together to perform such functions as contraction and secretion, which are characteristic of platelet activation.

Table 18.1 Some contents of platelet granules

Dense (δ) granules	α Granules	Lysosomal vesicles
ATP	PF4	Galactosidases
ADP	β-Thromboglobulin	Fucosidases
Calcium	Fibrinogen	Hexosaminidase
Serotonin	Factor V	Glucuronidase
Pyrophosphate	Thrombospondin	Cathepsin
P selectin (CD62P)	Fibronectin	Glycohydrolases
Transforming growth factor-beta (1)	PDGF	+ others
Catecholamines (epinephrine/norepinephrine)	PAI-1
GDP/GTP	Histidine-rich glycoprotein
	α₂ Macroglobulin
	Plasmin inhibitor
	P selectin (CD62)

ADP, adenosine 5′-diphosphate; ATP, adenosine 5′-triphosphate; GDP, guanosine 5′-diphosphate; GTP, guanosine 5′-triphosphate; PAI-1, plasminogen activator inhibitor-1; PDGF, platelet-derived growth factor; PF4, platelet factor 4.

The platelet membrane is the site of interaction with the plasma environment and with the damaged vessel wall. It consists of phospholipids, cholesterol, glycolipids and at least nine glycoproteins, named GPI to GPIX. The membrane phospholipids are asymmetrically distributed, with sphingomyelin and phosphatidylcholine predominating in the outer leaflet and phosphatidyl-ethanolamine, -inositol and -serine in the inner leaflet. After platelet activation the membrane also expresses binding sites for several coagulation proteins, including factor XI and factor VIII.

The contractile system of the platelet consists of the dense microtubular system and the circumferential microfilaments, which maintain the disc shape. Actin is the main constituent of the contractile system, but myosin and a regulatory calcium-binding protein, calmodulin, are also present.

Platelet Function in the Haemostatic Process

The main steps in platelet function ⁷ are adhesion, activation with shape change and aggregation. When the vessel wall is damaged, the subendothelial structures, including basement membrane, collagen and microfibrils, are exposed. VWF binds to collagen and microfibrils and then captures platelets via initial binding to platelet GPIb, resulting in an initial monolayer of adhering platelets. Binding via GPIb initiates activation of the platelet via a G-protein mechanism. Once activated, platelets immediately change shape from a disc to a tiny sphere with numerous projecting pseudopods. After adhesion of a single layer of platelets to the exposed subendothelium, platelets stick to one another to form aggregates. Fibrinogen, fibronectin, further VWF released from platelets and the glycoprotein Ib-IX and IIbIIIa complexes are essential at this stage to increase the cell-to-cell contact and facilitate aggregation. Certain substances (agonists) react with specific platelet membrane receptors to promote platelet aggregation and further activation. The agonists include exposed collagen fibres, ADP, thrombin, adrenaline, (epinephrine) serotonin and certain arachidonic acid metabolites including TXA₂. In areas of non-linear blood flow, such as may occur at the site of an injury, locally damaged red cells release ADP, which further activates platelets.

Platelet Aggregation

Platelet aggregation may occur by at least two independent but closely linked pathways. The first pathway involves arachidonic acid metabolism. Activation of phospholipase enzymes (PLA₂) releases free arachidonic acid from membrane phospholipids (phosphatidyl choline). About 50% of free arachidonic acid is converted by a lipo-oxygenase enzyme to a series of products including leucotrienes, which are important chemoattractants of white cells. The remaining 50% of arachidonic acid is converted by the enzyme cyclooxygenase into labile cyclic endoperoxides, most of which are in turn converted by thromboxane synthetase into TXA₂. TXA₂ has profound biological effects, causing secondary platelet granule release and local vasoconstriction, as well as further local platelet aggregation via the second pathway below. It exerts these effects by raising intracellular cytoplasmic free calcium concentration and binding to specific granule receptors. TXA₂ is very labile with a half-life of <1 min before it is degraded into the inactive thromboxane B₂ (TXB₂) and malonyldialdehyde.

The second pathway of activation and aggregation can proceed completely independently from the first one: various platelet agonists, including thrombin, TXA₂ and collagen, bind to receptors and via a G-protein mechanism, activate phospholipase C. This generates diacylglycerol and inositol triphosphate, which in turn activate protein kinase C and elevate intracellular calcium, respectively. Calcium is released from the dense tubular system to form complexes with calmodulin; this complex and the free calcium act as coenzymes for the release reaction, for the activation of different regulatory proteins and of actin and myosin and the contractile system and also for the liberation of arachidonic acid from membrane phospholipids and the generation of TXA₂.

The aggregating platelets join together into loose reversible aggregates, but after the release reaction of the platelet granules, larger, firmer aggregates form. Changes in the platelet membrane configuration now occur; ‘flip-flop’ rearrangement of the surface brings the negatively charged phosphatidyl-serine and -inositol on to the outer leaflet, thus generating platelet factor 3 (procoagulant) activity. At the same time specific receptors for various coagulation factors are exposed on the platelet surface and help coordinate the assembly of the enzymatic complexes of the coagulation system. Local generation of thrombin will then further activate platelets.

Platelets are not activated if in contact with healthy endothelial cells. The ‘non-thrombogenicity’ of the endothelium is the result of a combination of control mechanisms exerted by the endothelial cell: synthesis of prostacyclin, capacity to bind thrombin and activate the PC system, ability to inactivate vasoactive substances and so on. Prostacyclin released locally binds to specific platelet membrane receptors and then activates the membrane-bound adenylate cyclase (producing cyclic adenosine monophosphate or cAMP). cAMP inhibits platelet aggregation by inhibiting arachidonic acid metabolism and the release of free cytoplasmic calcium ions.

Thus platelets have at least three roles in haemostasis:

1. Adhesion and aggregation forming the primary haemostatic plug

2. Release of platelet activating and procoagulant molecules

3. Provision of a procoagulant surface for the reactions of the coagulation system.

Blood Coagulation

The central event in the coagulation pathways ⁸ is the production of thrombin, which acts upon fibrinogen to produce fibrin and thus the fibrin clot. This clot is further strengthened by the crosslinking action of factor XIII, which itself is activated by thrombin. The two commonly used coagulation tests, the activated partial thromboplastin time (APTT) and the prothrombin time (PT), have been used historically to define two pathways of coagulation activation: the intrinsic and extrinsic paths, respectively. However, this bears only a limited relationship to the way coagulation is activated in vivo. For example, deficiencies of factor XII or of factor VIII both produce marked prolongation of the APTT, but only deficiency of the latter is associated with a haemorrhagic tendency. Moreover, there is considerable evidence that activation of factor IX (intrinsic pathway) by factor VIIa (extrinsic pathway) is crucial to establishing coagulation after an initial stimulus has been provided by factor VIIa-tissue factor (TF) activation of factor X (Fig. 18.1).⁸

Figure 18.1 Schematic representation of the coagulation network. The major interactions are shown by the bold arrows. Inhibitory factors and clot limiting mechanisms are not shown. HMWK, high molecular weight kininogen; PL, phospholipid; TFPI, tissue factor pathway inhibitor.

Investigation of the coagulation system centres on the coagulation factors, but the activity of these proteins is also greatly dependent on specific surface receptors and phospholipids largely presented on the surface of platelets and also by activated endothelium. The necessity for calcium in many of these reactions is frequently used to control their activity in vitro. The various factors are described in the following sections, as far as possible in their functional groups; their properties are detailed in Table 18.2.

Table 18.2 The coagulation factors

The Contact Activation System

The contact activation system ⁹ comprises factor XII (Hageman factor), high molecular weight kininogen (HMWK) (Fitzgerald factor) and prekallikrein/kallikrein (Fletcher factor). As mentioned earlier, these factors are not essential for haemostasis in vivo. Important activities are to activate the fibrinolytic system, to activate the complement system and to generate vasoactive peptides: in particular, bradykinin is released from HMWK by prekallikrein or FXIIa cleavage. Kallikrein and factor XIIa also function as chemoattractants for neutrophils. The contact activation system also has some inhibitory effect on thrombin activation of platelets and prevents cell binding to endothelium. Recent evidence implicates the contact system in thrombosis via activation by polyphosphate released from platelets.¹⁰

When bound to a negatively charged surface in vitro, factor XII and prekallikrein are able to reciprocally activate one another by limited proteolysis, but the initiating event is not clear. It may be that a conformational change in factor XII on binding results in limited autoactivation that triggers the process. HMWK acts as a (zinc-dependent) cofactor by facilitating the attachment of prekallikrein and factor XI, with which it circulates in a complex, to the negatively charged surface. It has been shown in in vitro studies that platelets or endothelial cells can provide the necessary negatively charged surface for this mechanism and also possess specific receptors for factor XI. The contact system can activate fibrinolysis by a number of mechanisms: plasminogen cleavage, urokinase plasminogen activator (uPA) activation and tissue plasminogen activator (tPA) release. Most importantly from the laboratory point of view, the contact activation system results in the generation of factor XIIa, which is able to activate factor XI, thus initiating the coagulation cascade of the intrinsic pathway.

Fibrinogen

Fibrinogen ¹¹ is a large dimeric protein, each half consisting of three polypeptides named Aα, Bβ and γ held together by 12 disulphide bonds. The two monomers are joined together by a further three disulphide bonds. A variant γ chain denoted γ′ is produced by a variation in messenger RNA splicing. In the process a platelet binding site is lost and high-affinity binding sites for FXIII and thrombin are gained. The γ′ variant constitutes approximately 10% of plasma fibrinogen. A less common (<2%) γ chain variant ‘γE’ is also produced by splice variation. Fibrinogen is also found in platelets, but the bulk of this is derived from glycoprotein IIbIIIa-mediated endocytosis of plasma fibrinogen, which is then stored in alpha granules, rather than synthesis by megakaryocytes. Fibrin is formed from fibrinogen by thrombin cleavage releasing the A and B peptides from fibrinogen. This results in fibrin monomers that then associate and precipitate forming a polymer that is the visible clot. The central E domain exposed by thrombin cleavage binds with a complementary region on the outer or D domain of another monomer. The monomers thus assemble into a staggered overlapping two-stranded fibril. More complex interactions subsequently lead to branched and thickened fibre formation, making a complex mesh that binds and stabilizes the primary platelet plug.

Factor XIII

The initial fibrin clot is held together by non-covalent interactions and can be deformed and resolubilized. Factor XIII, which is also activated by thrombin, is able to covalently crosslink these fibrin monomers. Factor XIII is a transglutaminase that joins a glutamine residue on one chain to a lysine on an adjacent chain. This loss of resolubility is the basis of the screening test for factor XIII deficiency.

Inhibitors of Coagulation

A number of mechanisms exist to ensure that the production of the fibrin clot is limited to the site of injury and is not allowed to propagate indefinitely.^12,¹³ First, there are a number of proteins that bind to and inactivate the enzymes of the coagulation cascade. Probably the first of these to become active is TFPI, which rapidly quenches the factor VIIa–TF complex that initiates coagulation. It does this by combining first with factor Xa, so that further propagation of coagulation is dependent on the small amount of thrombin that has been generated during initiation being sufficient to activate the intrinsic pathway.

The principal physiological inactivator of thrombin is antithrombin (AT, formerly ATIII), which belongs to the serpin group of proteins. This binds to factor IIa forming an inactive thrombin–antithrombin complex (TAT), which is subsequently cleared from the circulation by the liver. This process is greatly enhanced by the presence of heparin or vessel wall heparan. AT is responsible for approximately 60% of thrombin-inactivating capacity in the plasma; the remainder is provided by heparin cofactor II and less specific inhibitors such as α2 macroglobulin. AT is also capable of inactivating factors X, IX, XI and XII but to lesser degrees than thrombin.

As thrombin spreads away from the area of damage it is also bound by thrombomodulin on the surface of endothelial cells. In this way it is changed from a primarily procoagulant protein to an anticoagulant one. Although remaining available for binding to AT, thrombin bound to thrombomodulin no longer cleaves fibrinogen. It now has a greatly enhanced preference for PC as a substrate. PC is presented to the thrombin–thrombomodulin complex by the endothelial protein C receptor (EPCR) and when activated by thrombin cleavage acts to limit and arrest coagulation by inactivating factors Va and VIIIa. This action is further enhanced by its cofactor, protein S, which does not require prior activation. The role of EPCR is particularly important in larger vessels, where the effective concentration of thrombomodulin is low. PC is subsequently inactivated by its own specific inhibitor.

The Fibrinolytic System

The deposition of fibrin and its removal are regulated by the fibrinolytic system.¹⁴ Although this is a complex multicomponent system with many activators and inhibitors, it centres around the fibrinogen- and fibrin-cleaving enzyme plasmin. Plasmin circulates in its inactive precursor form, plasminogen, which is activated by proteolytic cleavage. The principal plasminogen activator (PA) in humans is tissue plasminogen activator (tPA), which is another serine protease. tPA and plasminogen are both able to bind to fibrin via the amino acid lysine. Binding to fibrin brings tPA and plasminogen into close proximity so that the rate of plasminogen activation is markedly increased and thus plasmin is generated preferentially at its site of action and not free in plasma. The second important physiological PA in humans is called urokinase (uPA). This single chain molecule (scu-PA or pro-urokinase) is activated by plasmin or kallikrein to a two-chain derivative (tcu-PA), which is not fibrin-specific in its action. However, the extent to which this is important in vivo is not clear and the identification of cell surface receptors for uPA suggests that its primary role may be extravascular. The contact activation system also appears to generate some plasminogen activation via factor XIIa and bradykinin-stimulated release of tPA. The degradation products released by the action of plasmin on fibrin are of diagnostic use and are discussed later in this chapter. The activation of plasmin on fibrin is restricted by the action of a carboxypeptidase, which removes the amino terminal lysine residues to which plasminogen and tPA bind. This carboxypeptidase is activated by thrombomodulin-bound thrombin and is referred to as thrombin-activated fibrinolysis inhibitor (TAFI).

PAI-1 (plasminogen activator inhibitor-1) is a potent inhibitor of tPA, produced by endothelial cells, hepatocytes, platelets and placenta. Levels in plasma are highly variable. It is a member of the serpin family and is active against tPA and tcu-PA. A second inhibitor PAI-2 has also been identified, originally from human placenta, but its role and importance are not yet established.

The main physiological inhibitor of plasmin in plasma is plasmin inhibitor (α2-antiplasmin), which inhibits plasmin function by forming a 1:1 complex (plasmin–antiplasmin complex, PAP). This reaction in free solution is extremely rapid but depends on the availability of free lysine-binding sites on the plasmin. Thus, fibrin-bound plasmin in the clot is not accessible to the inhibitor. Deficiencies of the fibrinolytic system are rare but have sometimes been associated with a tendency to thrombosis or haemorrhage.

General approach to investigation of haemostasis

This section begins with some general points regarding the clinical and laboratory approach to the investigation of haemostasis. Following this, the basic or first-line screening tests of haemostasis are described. These tests are generally used as the first step in investigation of an acutely bleeding patient, a person with a suspected bleeding tendency or as a precaution before an invasive procedure is carried out. They have the virtue that they are easily performed and the patterns of abnormalities obtained point clearly to the appropriate next set of investigations. It should be remembered, however, that these tests examine only a portion of the haemostatic mechanism and have limited sensitivity for the presence of significant bleeding diatheses such as von Willebrand disease (VWD) or disorders of platelets or vessels. Hence a normal ‘clotting screen’ should not be taken to mean that haemostasis is normal.¹⁵

Clinical Approach

The investigation of a suspected bleeding tendency may begin from three different points:

1. Investigating a clinically suspected bleeding tendency. The investigation properly begins with the bleeding history, which may suggest an acquired or congenital disorder of primary or secondary haemostasis. If the bleeding history or family history is significant, appropriate specific tests and assays should be performed, notwithstanding the results of screening tests such as the PT, APTT and so on. Considerable effort has been put into defining those aspects of clinical history that predict a significant bleeding disorder and bleeding state questionnaires are now available.¹⁶

2. Following up an abnormal first-line test. The abnormalities already detected will determine the appropriate further investigations (discussed later).

3. Investigation of acute haemostatic failure. This is often required in the context of an acutely ill or postoperative patient. Investigations are therefore directed toward detecting disseminated intravascular coagulation (DIC) or a previously undetected coagulation defect (congenital or acquired). The availability of a normal premorbid coagulation screen and further questioning to determine a bleeding history can be extremely useful in this respect.

In all cases, comprehensive clinical evaluation, including the patient’s history, the family history and the family tree, as well as the details of the site, frequency and the character of haemorrhagic manifestations (purpura, bruising, large haematomata, haemarthroses, etc.), are required to establish a definitive diagnosis. If considered in conjunction with laboratory results, they will help avoid misinterpretation. It is also desirable to undertake a series of screening tests before proceeding to more specific tests. The results of the screening investigations, taken in conjunction with clinical information, usually point to the appropriate additional procedure.

Principles of Laboratory Analysis

It is worth remembering that the tests of coagulation performed in the laboratory are attempts to mimic in vitro processes that normally occur in vivo. Not surprisingly, this may give rise to misleading results. One of the most striking is the gross prolongation of the APTT in complete factor XII deficiency in the absence of any bleeding tendency. Similarly, the amount of factor VII required to produce a normal PT is greatly in excess of the amount required for normal haemostasis. Conversely, normal screening tests do not necessarily imply that the patient has entirely normal haemostasis.

The more detailed investigations of coagulation proteins also require caution in their interpretation depending on the type of assay performed. These can be divided into three principal categories, as described in the following sections.

Immunological

Immunological tests include immuno-diffusion, immuno-electrophoresis, radioimmunometric assays, latex agglutination (immunoturbidimetric) tests and tests using enzyme-linked immunosorbent assays (ELISA). Fundamentally, all these tests rely on the recognition of the protein in question by polyclonal or monoclonal antibodies. Polyclonal antibodies lack specificity but provide relatively high sensitivity, whereas monoclonal antibodies are highly specific but produce relatively low levels of antigen binding. Immunological assays are often easy to perform, particularly convenient for large batches and can be bought as kits with standardized controls. The obvious drawback of these assays is that they may tell you nothing about the functional capacity of the antigen detected. If possible they should always be carried out in parallel with a functional assay.

With advances in automation, latex agglutination kits are becoming more popular and replacing the more established ELISA assays. Latex microparticles are coated with antibodies specific for the antigen to be determined. When the latex suspension is mixed with plasma an antigen–antibody reaction takes place, leading to the agglutination of the latex microparticles. Agglutination leads to an increase in turbidity of the reaction medium and this increase in turbidity is measured photometrically as an increase in absorbance. Usually the wavelength used for latex assays is 405 nm, although for some assays a wavelength of 540 or 800 nm is used. Instrument-specific application sheets should be followed for each kit. This type of assay is referred to as immunoturbidimetric. Do not freeze latex particles because this will lead to irreversible clumping. An occasional problem with latex agglutination assays is interference from rheumatoid factor or other autoantibodies. These may cause agglutination and overestimation of the protein under assay. It is then preferable to resort to an ELISA assay.

Assays using chromogenic peptide substrates (amidolytic assays)

The serine proteases of the coagulation cascade have narrow substrate specificities.¹⁷ It is possible to synthesize a short peptide specific for each enzyme that has a dye (p-nitroaniline, p-NA) attached to the terminal amino acid. When the synthetic peptide reacts with the specific enzyme, the dye is released and the rate of its release or the total amount released can be measured photometrically. This gives a measure of the enzyme activity present. Chromogenic substrate assays can be classified into direct and indirect assays. Direct assays can be further subclassified into primary assays, in which a substrate specific for the enzyme to be measured is used, and secondary assays, in which the enzyme or proenzyme measured is used to activate a second protease for which a specific substrate is available. Specific substrates are available for many coagulation enzymes. However, the substrate specificity is not absolute and most kits include inhibitors of other enzymes capable of cleaving the substrate to improve specificity. Indirect assays are used to measure naturally occurring inhibitors and some platelet factors.¹⁵

It should be remembered that the measurement of amidolytic activity is not the same as the measurement of biological activity in a coagulation assay and in some cases may not accurately reflect this. This is particularly important when dealing with the molecular variants of various coagulation factors. The assays can be automated, carried out in a microtitre plate or in a tube when a spectrophotometer is used to measure the intensity of the colour development.

Coagulation assays

Coagulation assays are functional bioassays and rely on comparison with a control or standard preparation with a known level of activity. In the one-stage system optimal amounts of all the clotting factors are present except the one to be determined, which should be as near to nil as possible. The best one-stage system is provided by a substrate plasma obtained either from a patient with severe congenital deficiency or artificially depleted by immuno-adsorption. The principles of bioassay, its standardization and its limitations are considered in detail on p. 417.

Coagulation techniques are also used in mixing tests to identify a missing factor in an emergency or to identify and estimate quantitatively an inhibitor or anticoagulant. The advantage of this type of assay is that it most closely approximates the activity in vivo of the factor in question. However, they can be technically more difficult to perform than the other types described earlier and their susceptibility to interference from other plasma components has a detrimental effect on their accuracy and precision.

Other Assays

Other assays include measurement of coagulation factors using snake venoms, assay of ristocetin cofactor and the clot solubility test for factor XIII. DNA analysis is becoming more useful and more prevalent in coagulation. However, this requires entirely different equipment and techniques (see Chapter 8).

Notes on equipment

Waterbaths

A 37°C waterbath is required for manual coagulation tests, incubation steps and the rapid thawing of frozen specimens. Waterbaths set at 37°C should vary by no more than ± 0.5°C because slight variation in temperature will markedly affect the speed of clotting reactions. A waterbath with plastic or glass sides is preferable and some type of cross-illumination helps to determine the exact time of formation and appearance of the fibrin clot. Check that the temperature is 37°C before and during use. Distilled water should be used to fill the waterbath and maintain the water level.

Refrigerators and Freezers

Check that the temperature has not been out of the acceptable range of 4 ± 2°C for refrigerators and −20 ± 2°C or −80 ± 2°C (as applicable) for freezers, rechecking during the day. Records must be kept.

Centrifuges

Check to ensure each machine is clean before and after use. Also do a visual inspection of rotors, buckets and liners for corrosion and cracks. Thorough maintenance records should be kept.

Reagents and Buffers

Attention must be paid to the age and condition of solutions. This is particularly important with the calcium chloride solution. Whenever a solution is prepared it should be correctly labelled and dated. Buffers should be inspected for bacterial growth before use: contamination with microorganisms can cause errors and assay failures as a result of the release of enzymes and other active biological substances into solution. Azide may be added as a preservative to some buffers but should not be used in reagents for platelet studies or ELISA substrates. Chromogenic substrates should be reconstituted with sterile distilled water; contamination with bacterial enzymes may cause para nitro-aniline (pNA) release and yellow discoloration of the reagent. Records of batch numbers and expiry dates should be kept.

Plastic and Glass Tubes

For clotting tests, 75 × 10 mm glass rimless test tubes should be used. Plastic tubes should be used for sample dilutions, storage and reagent preparation.

Pipettes

A range of graduated glass (certified Class A) and automatic pipettes must be obtained. The latter should be accurate and durable. Fluids should not be drawn into the pipette barrels and acids should not be pipetted with instruments containing metal piston assemblies, which may become pitted or corroded. Attention to technique is vital because contamination of reagents with used pipette tips may occur, there may be errors of volume as a result of fluid on the exterior of the pipette tip or the manner of addition of a reagent may alter the results obtained. The amount of fluid drawn into the tip should be inspected visually with each pipetting procedure. Records of pipette accuracy and precision should be kept.

Stopwatches and Stopclocks

Stopclocks are useful for timing incubation periods of several minutes or more, but stopwatches that may be held in the hand, and controlled rapidly, should be used for measuring clotting times and for short incubations. At least four stopwatches are needed unless an automated coagulometer is used.

Automated Coagulation Analysers

A wide variety of automated and semi-automated coagulation analysers are available. The choice of analyser depends on predicted workload, repertoire and cost implications. A thorough evaluation of the current range of analysers is recommended.

Modern analysers distributed in the European Economic Area must be CE marked, certifying that the product has met EU consumer safety, health or environmental requirements.

If coagulation analysers are used, it is important to ensure that their temperature control and the mechanism for detecting the endpoint are functioning properly. Although such instruments reduce observer error when a large number of samples are tested, it is important to apply stringent quality control at all times to ensure accuracy and precision.

Evaluating and choosing an automated analyser

The purchasing or leasing of new equipment is a complicated process and the most important factors to be considered will vary from one laboratory to another.

Specification standards may be classified into Mandatory and Desirable. An example classification is shown below:

Desirable additional requirements

• Good reputation of supplier and satisfaction of other users

• Acceptable level of service, telephone support and availability of engineer support

• Availability of independent evaluation report

• Satisfactory performance of analyser and reagent combinations in an external quality assessment scheme

• Satisfactory results of an on-site evaluation

• Ease of implementation in laboratory with minimal necessary changes to existing practice

• Ease of implementation of future requirements and increase in workload

• Guaranteed supply of reagents and consumables

• Post-sales support, training and education.

The final decision is usually made after competitive tenders are submitted to ensure fairness to all relevant commercial firms and after achieving the lowest appropriate cost. The selection process should take into account the following cost implications:

• Capital bid cost or leasing/rental cost

• Cost of reagents and consumables per year

• Cost of maintenance contracts per year

• Cost of computer interfacing.

The extent to which each analyser fulfils the essential and desirable attributes can be scored according to their relative importance. Thus, for example, the specification standard should be weighted, 10 for the most important, 1 for the least important criterion; compliance with the specification standard should be given a mark: 5 as the best and 1 as the worst score.

A total score is then calculated as Weighting × Mark and summed for each analyser.

Safety

Each laboratory should have its own safety recommendations and procedures to follow in case of accident or contamination (see Chapter 24).

Pre-analytical variables including sample collection

Many misleading results in blood coagulation arise not from errors in testing but from carelessness in the pre-analytical phase. Ideally, the results of blood tests should accurately reflect the values in vivo.

When blood is withdrawn from a vessel, changes begin to take place in the components of blood coagulation. Some occur almost immediately, such as platelet activation and the initiation of the clotting mechanism dependent on surface contact.

It is essential to take precautions at this early stage to prevent, or at least minimize, in vitro changes by conforming to recommended criteria during collection and storage. These criteria, as described below, have been established by the Clinical and Laboratory Standards Institute (CLSI).

Collection of Venous Blood

Venous blood samples should be obtained whenever possible, even from the neonate. Capillary blood tests require modification of techniques, experienced operators and locally established normal ranges; they are not an easy alternative to tests on venous blood. All blood samples must be collected by personnel who are trained and experienced in the technique. Patients requiring venepuncture should be relaxed and in warm surroundings. Excessive stress and vigorous exercise cause changes in blood clotting and fibrinolysis. Stress and exercise will increase factor VIII, VWF and fibrinolysis.

Whenever possible, venous samples should be collected without a pressure cuff, allowing the blood to enter the syringe by continuous free flow or by the negative pressure from an evacuated tube (see p. 3). Venous occlusion causes haemoconcentration, increase of fibrinolytic activity, platelet release and activation of some clotting factors. In the majority of patients, however, light pressure using a tourniquet is required; this should be applied for the shortest possible time (e.g. <1 min). The venepuncture must be ‘clean’; blood samples from an indwelling line or catheter should not be used for tests of haemostasis because they are prone to dilution and heparin contamination.

To minimize the effects of contact activation, good-quality plastic or polypropylene syringes should be used. If glass blood containers are used, they should be evenly and adequately coated with silicon.

The blood is thoroughly mixed with the anticoagulant by inverting the container several times. The samples should be brought to the laboratory as soon as possible. If urgent fibrinolysis tests are contemplated, the blood samples should be kept on crushed ice until delivered to the laboratory. Assays of tPA and of PA1-1 antigen are preferably performed on samples taken into trisodium citrate to prevent continued tPA–PA1-1 binding (see p. 621).

If an evacuated tube system is used for collecting samples for different tests, the coagulation sample should be the second or third tube obtained.

Patient identification is of utmost importance. Care must be taken in labelling the patient sample both at the bedside and within the laboratory.

Blood Sample Anticoagulation

The most commonly used anticoagulant for coagulation samples is trisodium citrate. A 32 g/1 (0.109 M) solution (see p. 621) is recommended. Other anticoagulants, including oxalate, heparin and ethylenediaminetetra-acetic acid (EDTA), are unacceptable. The labile factors (factors V and VIII) are unstable in oxalate, whereas heparin and EDTA directly inhibit the coagulation process and interfere with endpoint determinations. Additional benefits of trisodium citrate are that the calcium ion is neutralized more rapidly in citrate and APTT tests are more sensitive to the presence of heparin.

For routine blood coagulation testing, 9 volumes of blood are added to 1 volume of anticoagulant (i.e. 0.5 ml of anticoagulant for a 5 ml specimen). When the haematocrit is abnormal with either severe anaemia or polycythaemia, the blood:citrate ratio should be adjusted.¹⁸ For a 5 ml specimen (total), the amount of citrate should be as follows:

Haematocrit	Citrate (ml)
0.20	0.70
0.25	0.65
0.30	0.61
0.55	0.39
0.60	0.36
0.65	0.30
0.70	0.26

Transportation to the Laboratory

An efficient and regular collection service is necessary. It is important that samples are delivered as quickly as possible to prevent deterioration of the labile clotting factors such as factors V and VIII. Automated systems can facilitate rapid delivery, but should be avoided when platelet function tests are to be performed because the associated trauma may affect results. For certain investigations it is necessary for the samples to be placed on ice once taken and delivered immediately to the laboratory.

Centrifugation: Preparation of Platelet-Poor Plasma

Most routine coagulation investigations are performed on platelet-poor plasma (PPP), which is prepared by centrifugation at 2000 g for 15 min at 4°C (approx. 4000 rev/min in a standard bench cooling centrifuge). The sample should be kept at room temperature if it is to be used for PT tests, lupus anticoagulant (LAC) or factor VII assays and it should be kept at 4°C for other assays; the testing should preferably be completed within 2 h of collection. Care must be taken not to disturb the buffy coat layer when removing the PPP.

Samples for platelet function testing, LAC and the activated PC resistance (APCR) test should not be centrifuged at 4°C. These samples should be prepared by centrifugation at room temperature to prevent activation of platelets and release of platelet contents such as phospholipid and factor V. For LAC testing and APCR it is very important that the number of platelets and the amount of platelet debris in the samples are minimized. The platelet count should be below 10 × 10⁹/l. This is best achieved by double centrifugation or filtration of the plasma through a 0.2 μm filter.

Storage of Plasma and Sample Thawing

Some tests such as the PT and APTT are carried out on fresh samples. Certain coagulation assays, unless urgently required, can be performed in batches at a later date on deep frozen plasma. Storage of small aliquots of samples in liquid nitrogen (−196°C) is the optimum, although samples may be frozen at −40°C or −80°C for several weeks without significant loss of most haemostatic activities. Gentle but thorough mixing of samples is essential after thawing and before testing. Once thawed, the sample should never be refrozen.

Some Common ‘Technical’ Errors

A false abnormality of the clotting time may occur in the following situations:

1. Faulty collection of the sample, resulting in it undergoing partial clotting

2. Underfilling or overfilling of the bottle or high or low haematocrit (can cause the volume of citrate in relation to the plasma volume to be incorrect; see above)

3. An unsuitable anticoagulant, such as EDTA, used in collecting the sample

4. Collection of blood through a line that has at some stage been in contact with heparin

5. Contamination of the kaolin/platelet substitute reagent with a trace of thromboplastin

6. Undue delay in sample analysis

7. Use of inaccurate pipettes (documented proof of pipette calibration is essential)

8. Machine malfunction

9. Incorrect waterbath temperature

10. Calcium chloride at incorrect concentration or not freshly prepared.

Calibration and Quality Control

Reference Standard (Calibrator)

International (WHO) and national standards are available for a number of coagulation factors (see p. 589). For diagnostic tests it is necessary to test a calibrated normal reference preparation alongside the patients’ plasmas.

Because the concentration of some coagulation factors may vary as much as fourfold in different normal plasma samples, it is inadvisable to use plasma from any one person to represent 100% clotting activity. The larger the number of donors in the pool, the more likely the pool clotting activity will be 100% or 100 iu/dl. A suggested minimum for the normal pool is 20 donors. It is preferable to use a calibrated reference plasma for routine use with each assay. If this is not possible, then a locally prepared normal pool can be used provided it is itself calibrated against a reference preparation.

Calibration of Standard Pools and Suggested Calibration Procedure

Whenever possible, the normal pool should be calibrated as described in the following against a freeze-dried reference material already calibrated against the international standard. The reference material may be a national standard (e.g. National Institute for Biological Standards and Control) or a commercial standard. In the absence of reference materials the laboratory should obtain as large a normal pool as possible and assign it a value of 100 iu/dl.

The most important principle of calibration is repetition to minimize possible errors at each stage of calibration. It is necessary to carry out at least four independent assays and preferably six. An independent assay is an assay for which a new ampoule of standard is opened, or if a freeze-dried standard is not available, for which a new set of dilutions are prepared from frozen previous reference plasma. Each plasma must be tested in duplicate; two replicate assays should be carried out each day and the procedure should be repeated on at least 4 days (four independent assays). Whenever possible more than one operator should be involved.

Comparison should always be made with the previous normal pool. The potency of the new normal pool is calculated for each replicate assay on each day and an overall mean value is calculated. This calibration also enables an assessment of the precision of the method used.

Control Plasma

Controls are included alongside patient samples in a batch of tests. Inclusion of both normal and abnormal controls will enable detection of non-linearity in the standard curve. Whereas a reference standard (calibrator) is used for accuracy, controls are used for precision. Precision control, the recording of the day-to-day variation in control values, is an important procedure in laboratory coagulation. Participation in an external assessment scheme (see p. 594) is also important to ensure inter-laboratory harmonization. The use of lyophilized reference standard and control plasmas has become widespread, whereas locally calibrated standard pools are used especially in under-resourced countries. The results of participation in external qualitycontrol schemes require careful attention. The large number of different reagents, substrate plasmas, reference preparations and analysers available makes comparison of like with like difficult. Ideally all combinations should give similar results, but this is often not the case and the results should be used to carefully choose the combination used.

A control must be stable and homogeneous; the exact potency is not important, although the approximate value should be known to select a preparation at the upper or lower limit of the normal reference range.

Fresh control blood is required for procedures such as platelet aggregation and should be obtained from ‘normal’ healthy subjects. Fresh controls should be prepared in exactly the same way as the patient sample. Normal and abnormal controls are usually obtained from commercial companies.

Variability of Coagulation Assays

Within a laboratory, variability is most commonly the result of a dilution error, differences in the composition of reagents, failure to take the time-trend into account and differences in experience and technique between operators. A coefficient of variation of 15–20% is not uncommon for factor VIII assays but should be much lower for antigenic assays.

Variability between laboratories is much higher. Apart from the factors described for the within-laboratory variability, there is the major effect of differences in methods and in the composition of reagents. Comparability between laboratories improves if standardized reagents are used.

The unavoidable variability associated with coagulation assays makes the use of reliable reference materials imperative.

Performance of Coagulation Tests

Handling of Samples and Reagents

All plasma samples should be kept in plastic or siliconized glass tubes and placed on melting ice or at 4°C until used, except when cold activation of factor VII and platelets is to be avoided, in which case the plasma is kept at room temperature. All pipetting should be performed using disposable plastic pipettes or autodiluter pipette tips. The actual clotting tests are performed at 37°C in new round-bottom glass tubes of standard size (10 or 12 mm external diameter). Ideally, all glassware should be disposable. If the tubes have to be reused, scrupulous cleaning using chromic acid and a detergent such as 2% Decon 90 is essential.

Eliminating a Time Trend

The potential instability of biological reagents used in tests of haemostasis makes it desirable to arrange results so as to reduce time-related errors. Thus, if there is a significant length of time between the test with the patient’s plasma and the test with the control sample, the difference may be the result of the deterioration of one or more of the reagents or of the plasma itself rather than a true defect or deficiency. In the simplest case, if there are two samples A and B, the readings should be carried out in the order A₁, B₁, B₂, A₂. Additional specimens are allowed for by inserting further letters into the design.

Assay Monitoring and Endpoint Detection

Manual Methods

Detecting clot formation as the endpoint depends to some extent on the rate of its formation: the shorter the clotting time the more opaque is the clot and the easier it is to detect. A slowly forming clot may appear as mere fibrin wisps, which are difficult to detect by eye or machine. In manual work, the observer must try to adopt a uniform convention in selecting the moment in clot formation that will be accepted as the endpoint. It is also important to ensure that the tube can be watched with its lower part under the water or while being quickly dipped in and out so as to avoid cooling and a slowing down of the clot formation. Bubbles also make the determination of the endpoint difficult.

Manual clotting techniques are still used in WHO calibration schemes and therefore should be viewed as an essential skill, despite the ever-increasing reliance on automation. It is worth remembering that not all results produced by an automated analyser are correct; dubious or inconsistent results should be checked manually.

In instrumental work the coagulometer must be shown to detect long clotting times reliably and reproducibly. The various coagulometers available have different means of detecting the endpoint, which may make comparison of results difficult. Some commonly used techniques are as follows.

Electromechanical

Impedance, steel ball

The sample cuvette rotates and a steel ball (e.g. Amelung KC4A) remains stationary in a magnetic field until the formation of fibrin strands around the ball produces movement. This is detected by a change in the magnetic field and the coagulation time is recorded.

A steel ball (e.g. Axis-Shield Group; Thrombotrack) rotates under the influence of a magnet until the formation of fibrin strands around the ball stops it rotating. This is detected by a sensor and the coagulation time is recorded.

Photo-Optical Analysis

Scattered light detection for clotting assays (660 nm)

The turbidity during the formation of a fibrin clot is measured as an increase in scattered light intensity when exposed to light at a wavelength of 660 nm.

Transmitted light detection for chromogenic assays (405 nm, 575 nm, 800 nm)

Colour production leads to a change in light absorbance, which is detected as a change in transmitted light. Over time, the change in absorbance per minute is calculated (OD/min). Various wavelengths can be used such as 405 nm, 575 nm and 800 nm.

Transmitted light detection for immunoassays (405 nm, 575 nm, 800 nm)

The change in light absorbance caused by the antigen–antibody reaction is detected as the change in transmitted light. Over time, the change in absorbance per minute is calculated (OD/min). Some analysers detect light transmittance at multiple wavelengths between 395 and 710 nm.

Nephelometry

Nephelometry (IL ACL analysers) is the determination of the intensity of light scatter using a detector placed at right angles to the incident light path and detecting light of the same wavelength as the incident light. The procedure is particularly useful in measuring complexes of antigen and antibody produced by immunoprecipitation.

Photo-optical endpoint determination and analyses

A number of methods can be used to define the change in optical transmission that corresponds to the endpoint of the reaction.

Percentage Detection Method

After initiating the clotting reaction, the transmitted light is monitored and a baseline A/D value (bH) is determined for the reaction (bH = 0%) (Fig. 18.2). The reaction is then monitored until the clotting reaction is completed (dH = 100%). The time to an optionally set endpoint, usually 50%, is then determined. At this point the A/D value per unit time shows the greatest change and the fibrin monomer polymerization reaction rate is high. Detection based on this principle enables coagulation analysis to be more accurate at low fibrinogen concentrations in samples with low A/D values and those samples for which the initial amount of A/D value is higher than usual, such as lipaemic and haemolysed samples.

Figure 18.2 Endpoint determination: percentage detection method.

(Reproduced by permission of Sysmex UK Ltd.)

Rate Method

After the start of the reaction, the increase in absorbance per minute is monitored. At the predetermined end time, the final increase in absorbance per minute measurement is made. The rate of the absorbance increase per minute between these two time point measurements is calculated (Fig. 18.3). The calculated change in absorbance (dOD/min) is expressed as the raw data and used to construct a standard curve (i.e. there is no endpoint per se).

Figure 18.3 Analysis of reaction: rate method. This is used for chromogenic assays.

(Reproduced by permission of Sysmex UK Ltd.)

VLin Integral Method

The VLin integral method (Fig. 18.4) evaluates the absorbance per minute of an immunological reaction. This is monitored and mathematical analysis used to determine the peak rate of reaction (maximum velocity). Using this method allows for increase in analytical sensitivity, extended measuring range, reduced measurement time and improved antigen excess reliability when measuring an immunological reaction. The VLin integral evaluation method is used for immunological assays, including D-dimer and VWF antigen.

Figure 18.4 Analysis of reaction: VLin integral method. This is used for immunological assays.

(Reproduced by permission of Sysmex UK Ltd.)

Analysis Time Over

The ‘Analysis Time Over’ check (Fig. 18.5) is used to detect whether the reaction endpoint is correct. If the sample reaction end angle is greater than the permitted angle at maximum detection time, the result will be flagged with an ‘Analysis Time Over’ error. The situation occurs when testing samples with prolonged clotting times and satisfactory endpoint has not been reached by the end of the time allotted for analysis. When this occurs the following checks should be performed:

1. Check the sample for possible anticoagulant contamination, haemolysis, lipaemia, hyperbilirubinaemia or turbidity.

2. Verify delivery of sample and reagent.

3. Set the ‘Maximum Reading Time’ to a longer time and reanalyse the sample.

4. If reanalysis of the sample results in a numeric value without an error flag, the result can be reported.

5. If reanalysis gives an ‘Analysis Time Over’ message again, the sample may not be capable of forming a firm clot. In these situations, the clotting time must be checked manually.