Keywordsplatelets, liver, coagulation, HELLP syndrome, hemolysis, thrombophilia
Editors’ comment: This chapter of the fourth edition of Chesley’s textbook has been updated. In order to do so, the editors have enlisted the aid of Dr. Keith McCrae who has special expertise and clinical and research interests in platelets and their function in pregnancy and preeclampsia. As with the description of brain and liver pathology with the preeclampsia syndrome, we chose to discuss alterations in liver pathology and function in their clinical context rather than anatomically. Combined together, microangiopathic hemolysis, hepatocellular disruption, and thrombocytopenia make up the HELLP syndrome, which is discussed in this chapter as well as in Chapter 2 , Chapter 20 .
In his first edition, Chesley included a chapter entitled Disseminated Intravascular Coagulation and began by stating the contemporaneous thinking that this was a fundamental feature of preeclampsia–eclampsia. In his usual thorough fashion, he reviewed data that had accrued up to that time, and he concluded that there was evidence for slightly increased coagulation and fibrinolysis during normal pregnancy. He went on to say, however, that many women with severe preeclampsia and eclampsia show no detectable evidence of increased coagulation and fibrinolysis. He concluded that disseminated intravascular coagulation did not appear to be a fundamental feature of the disease.
Many of Chesley’s predecessors and contemporaries who espoused activation of intravascular coagulation with preeclampsia syndrome drew their conclusions from autopsy findings that undoubtedly led to some of these erroneous findings. Although observed as early as 1924, it had been proven by that time that platelet concentrations were decreased in some women with preeclampsia syndrome – especially severe cases that included those with eclampsia. In the first large study, Pritchard et al. reported a mean platelet count of 206,000/μL in 91 consecutive women with eclampsia. In a fourth of these, the platelet count was<150,000/μL, in 15% it was<100,000 μL, and in 3% it was<50,000/μL. From these and other studies, Chesley concluded that thrombocytopenia is a feature of the preeclampsia syndrome, but that it was not caused by consumptive coagulopathy.
Somewhat parallel to the coagulation story, it had been long known that severe preeclampsia and eclampsia were associated with gross and microscopical changes in the liver. A search for a serum analyte for hepatocellular necrosis came with documentation of elevated serum glutamic oxaloacetic transaminase (SGOT) levels. In his first edition chapter entitled The Liver , Chesley summarized 11 studies of SGOT measurements and he cited abnormal values in 84% of women with eclampsia, half of those with severe preeclampsia, and a fourth who had mild preeclampsia. And while hepatocellular damage is a known cause of coagulopathy, Chesley concluded that damage to the liver was generally not severe enough to cause significant liver dysfunction. But the link between thrombocytopenia and liver involvement characterized by elevated serum transaminase levels did evolve as a marker for the severity of preeclampsia. To call attention to this, Weinstein coined the term HELLP syndrome – H emolysis, E levated L iver enzymes, and L ow P latelets.
Thus coagulation, thrombocytopenia, and hepatic changes of the preeclampsia syndrome became accepted as interrelated. As with any review concerning preeclampsia, a major difficulty is the use of variable or imprecise criteria for its diagnosis, as discussed in Chapter 1 . This caveat must be considered even when comparing studies cited in this chapter.
Platelets are the smallest of the formed blood elements, with a diameter of 1.5 to 3 μm, a volume approximating 7 fL and a lifespan in vivo of 9 to 10 days. Platelets are extremely complex morphologically and biochemically, and have myriad functions. These small discoid elements consist of a plasma membrane, cytoskeletal elements, and several organelles, some of which communicate with the surface via an open canalicular system. Numerous membrane receptors serve to discharge platelet functions, the primary one being their adaptation to adhere to damaged blood vessels, with one another, and to stimulate thrombin generation, all of which generate a hemostatic plug – the clot.
Critical to platelet function are platelet surface receptors that bind adhesive glycoproteins. These receptors include the GPIb/IX/V complex that binds von Willebrand factor, the integrin GPIIb/IIIa (αIIbβ3) receptor which binds fibrinogen and von Willebrand factor, and GPIa/IIa, which binds collagen. Others receptors bind additional matrix glycoproteins, while the P-selectin receptor mediates interactions with leukocytes to incite a proinflammatory response.
Ligand binding by platelet cell surface receptors may induce platelet activation, through “outside-in” signaling. An idea of the complexity of this process comes from consideration of the multiple agonists shown in Table 17.1 . Various stimuli, including collagen, thrombin, serotonin, epinephrine, thromboxane A 2 , platelet-activating factor, and ADP, can stimulate platelet activation. When endothelium is disrupted, platelets adhere to exposed subendothelial collagen. This process requires von Willebrand factor and results in platelet shape change from a discoid to a tiny sphere with numerous fine filopodia or pseudopodia. Most platelets that accumulate at a site of injury do not adhere directly to subendothelial surfaces, but rather aggregate with each other, a process mediated primarily through the effects of platelet GPIIbIIIa, which adopts an active conformation capable of binding fibrinogen as a consequence of platelet activation. Inherent to the platelet activation process is secretion of the contents of dense granules and alpha granules that contain a variety of substances, such as some of those shown in Table 17.1 .
|Dense granules – ADP, ATP, GDP, Ca 2+ , Mg 2+ , serotonin, pyrophosphate|
|α-Granules – platelet-specific proteins (β-thromboglobulin family, platelet factor 4, multimerin)|
|Adhesive glycoproteins – fibrinogen, vWF, fibronectin, thrombospondin-1, vitronectin|
|Coagulation factors – factors V and X1, protein S|
|Mitogenic factors – PDGF, TGF-3, ECGF, EGF, IGF-1|
|Angiogenic factors – VEGF, PF4 inhibitor (α -PI), plasminogen-activator inhibitor-1 (PAI-1) Fibrinolytic inhibitors|
|Granule membrane-specific proteins – P-selectin (CD62P), CD63, GMP33 (thrombospondin fragment)|
|Others – proteases, interleukins, chemokines, inhibitors|
Platelets in Normal Pregnancies and with Preeclampsia
There are a number of normal pregnancy-induced changes that relate to platelets and their various functions. Some of these myriad changes in platelet numbers, morphology, and function are shown in Table 17.2 . Several large, population-based studies have demonstrated that in uncomplicated pregnancies, the platelet count decreases by about 10% by term. In one study, the mean platelet count in 6770 pregnant women near term was 213,000/μL, compared with 248,000/μL in nonpregnant control women. The 2.5th percentile in the pregnant group, used to define the lower limit of normal platelet concentrations, was 116,000/μL. The genesis of this decrease is not known for certain, but is likely to be related to the larger total blood volume as well as the expanded splenic volume, which may be as much as 35%.
|Factor||Preeclampsia vs. Normal Pregnancy||Comments|
|Concentration||Decreased||Dependent on severity and duration|
|Volume||Increased||Younger, larger platelets|
|Platelet activation in vivo|
|Beta-thromboglobulin||Increased (serum)||Associated with degranulation|
|Immune stimulation||Increased serum platelet-associated IgG|
|Cell adherence molecule expression||Increased||Increased expression anti-P-selectin, CD63, CD40+, CD60L|
|Thromboxane A 2||Urinary metabolites increased|
|Platelets in vivo|
|Aggregation||Decreased compared with increase of normal pregnancy||Reduced in response to ADP, arachidonic acid, vasopressin, and epinephrine|
|Release||Decreased||Reduced release of 5-hydroxytryptamine in response to epinephrine|
|Nitric oxide synthase||Decreased iNOS and peroxynitrite [NO(x)] – increased|
|Platelet second messengers|
|Intracellular free Ca 2+||Increased over normal pregnancy increase||Causes platelet activation|
|cAMP||Reduced cAMP platelet response to prostacyclin|
|Mg 2+ increases cAMP levels via prostacyclin|
|Angiotensin II||Normal levels compared with decreased levels in normotensive pregnancy||Angiotensin II enhances platelet aggregation with ADP and epinephrine|
Most studies have also observed that the platelet count decreases in preeclampsia/eclampsia. The variation of frequency and intensity of thrombocytopenia between studies may reflect differences between methods and/or equipment used for automated blood cell analysis. Up to half of women with preeclampsia develop thrombocytopenia, the extent of which is generally proportional to the severity of disease. The pathogenesis of preeclampsia-associated thrombocytopenia is likely multifactorial. The elevated levels of thromboxane A 2 metabolites in the urine of preeclamptic patients, as well as the increased plasma levels of the platelet α-granule proteins β-thromboglobulin and platelet factor 4, supports the argument that platelet activation contributes to accelerated platelet clearance in this disorder.
Although only the most severe cases of preeclampsia are associated with a coagulation profile suggestive of disseminated intravascular coagulation (DIC), the plasma of most patients with preeclampsia contains increased levels of thrombin-antithrombin complexes, and approximately 40% of these plasmas contain increased levels of fibrin D-dimers. As discussed on page 386, this suggests that there is at least subclinical activation of the coagulation system. Thus, increased generation of thrombin may be one mechanism that promotes platelet activation. Platelets may also be stimulated through contact with dysfunctional endothelium and/or exposed subendothelium underlying the injured placental vasculature. Platelet adhesion may be promoted by reduced levels of ADAMTS-13 as well as elevated levels and larger, more active multimers of von Willebrand factor (VWF) and other adhesive proteins such as cellular fibronectin.
Platelet volume increases normally across pregnancy. As shown in Table 17.2 , preeclampsia is associated with a further increase in mean platelet volume. This increase reflects a population of larger platelets thought to be the result of increased platelet consumption or destruction with a concomitantly increased proportion of young platelets. Other factors include complex changes in the pattern of platelet production and release by megakaryocytes. Optimal methods to investigate platelet lifespan require radiolabeling, which is prohibited in pregnancy. Using the method of platelet malondialdehyde production, disparate findings have been reported. Specifically, in their longitudinal study, Pekonen et al. did not find a significant reduction in platelet lifespan in preeclampsia. Conversely, Rakoczi et al. reported a significant decrease. The demonstration of a shorter platelet production time is consistent with a shorter platelet half-life. Importantly, the degree of thrombocytopenia is related to severity, and significant thrombocytopenia is usually associated with severe preeclampsia and eclampsia. These changes are found only with maternal platelets because, even with marked maternal thrombocytopenia with severe preeclampsia, neither cord blood nor fetal platelet counts are affected.
Redman et al. demonstrated that the platelet count fell at an early stage in the evolution of preeclampsia. Despite this, the absolute platelet count is of limited predictive or prognostic value. Of clinical significance are the findings of Leduc et al., who showed that in the absence of thrombocytopenia, women with severe preeclampsia do not have significant clotting abnormalities. And Barron et al. showed that a combination of a normal platelet count plus a normal serum lactate dehydrogenase level had a negative-predictive value of 100% for clinically significant clotting abnormalities in women with preeclampsia. Thus, studies to assess coagulation, viz. prothrombin and activated partial thromboplastin times and plasma fibrinogen concentration, can be reserved for women with platelet counts<100,000/μL.
This acronym defines a presumed preeclampsia variant manifest by hemolysis, elevated liver transaminases, and low platelets. Thrombocytopenia with HELLP syndrome is generally more severe than that encountered with uncomplicated preeclampsia syndrome. It has been reported that the rate of fall of the platelet count is a predictor of the eventual severity of HELLP, with women whose platelet counts decrease by>50,000/μL per day having a higher probability of developing moderate to severe thrombocytopenia with a platelet count<100,000/μL. There appears to be a correlation between the extent of thrombocytopenia and the degree of liver dysfunction in women with HELLP syndrome, as discussed on page 389. The platelet nadir is usually reached approximately 24 hours postpartum, with normalization occurring within 6–11 days. As with preeclampsia, the thrombocytopenia of the HELLP syndrome probably reflects a multifactorial pathogenesis. Thus, likely events operative include platelet activation by contact with damaged endothelium, platelet consumption secondary to thrombin generation, and microangiopathic hemolysis.
Platelet Activation In Vivo
Most studies have provided evidence that there is increased platelet activation across pregnancy. Moreover, this activation is increased further in women with preeclampsia ( Table 17.2 ). Circulating levels of factors stored within platelets reflect platelet activation – specifically, platelet aggregation and release of granule contents. Plasma levels of β-thromboglobulin, a platelet α-granule protein, are increased in normal pregnant women. Several studies have demonstrated higher plasma levels of β-thromboglobulin in preeclamptic women compared with normal pregnant controls. Janes and Goodall reported that these increased β-thromboglobulin concentrations were associated with degranulation, as evidenced by elevated levels of the lysosomal-granule membrane antigen, CD63. Socol et al. found that this measure of platelet α-granule release correlated with increasing levels of proteinuria and serum creatinine, suggesting a link between platelet activation and renal microvascular changes.
There are increased expressions of other platelet membrane antigens that signify activation. Platelets taken from preeclamptic women express increased levels of CD40L and its circulating soluble component, sCD40L. There are also increased levels of the antigens CD41- and CD62P+and their respective circulating microparticles along with platelet-monocyte aggregates.
Elevation of β-thromboglobulin levels precedes the clinical development of preeclampsia by at least 4 weeks. In contrast with normal pregnancy, levels of β-thromboglobulin did not correlate with increased fibrinopeptide A – a marker of thrombin generation – reported in preeclampsia. These findings suggest that mechanisms other than thrombin-mediated platelet stimulation are responsible for platelet activation. An immune mechanism may be a contributory factor. Burrows et al. reported increased serum levels of platelet-associated immunoglobulin G which correlated with disease severity. In a prospective study, Samuels et al. measured platelet-bound and circulating platelet-bindable immunoglobulin. They reported a higher frequency of abnormal platelet antiglobulin found in preeclamptic women compared with normotensive pregnant women. Alterations in platelet-bound immunoglobulins might be from the deposition of autoreactive antibodies or immune complexes caused by placental tissue antigens. Alternatively, platelet activation at sites of microvascular injury could lead to the externalization of IgG and other proteins in platelet α-granules.
Serum levels of platelet factor 4 (PF4), another α-granule protein, are not significantly elevated in preeclampsia. Because PF4 is cleared by binding to the endothelium rather than by renal excretion, a contribution of impaired renal function to the increased β-thromboglobulin levels cannot be excluded. Serotonin is also released when platelets aggregate, and Middelkoop et al. found decreased serotonin concentrations in platelet-poor plasma from preeclamptic women compared with levels in plasma from normal pregnant women. These results are consistent with platelet aggregation and consumption.
Janes and Goodall used whole blood flow cytometry to detect circulating activated platelets identified by bound fibrinogen or by CD63 antigen expression. They reported that activated platelets were detected prior to the development of preeclampsia. These findings were confirmed by Konijnenberg et al., who also demonstrated enhanced expression of anti-P-selectin, another marker of α-granule secretion.
The most reliable method of assessing in vivo thromboxane production is by measurement of urinary metabolites of thromboxane A. Urinary excretion of 2,3-dinor-thromboxane B and 11-dehydro-thromboxane B is increased in normal pregnancy. These urinary metabolites are further increased in women with preeclampsia compared with normotensive pregnant women, and may increase before clinical signs develop. These observations provide further evidence of increased platelet activation in preeclampsia and, moreover, indicate that thromboxane generation activates platelets.
Garzetti et al. studied platelet membranes and found increased fluidity and cholesterol concentration with preeclampsia. These changes are consistent with increased unsaturated fatty acid content. These latter compounds are both a substrate for lipid oxidation and participate in thromboxane formation. Increased thromboxane production may thus reflect altered platelet membranes in preeclampsia.
There may also be an activated-platelet interaction with the ADAMTS-13 metalloproteinase system that cleaves von Willebrand factor. Platelets interact with factor VIII to promote normal ADAMTS-13 action; however, disruption of this impairs vWF proteolysis both in vivo and in vitro . The effects of this on the normally deficient ADAMTS-13 state for pregnancy and the genesis of preeclampsia is speculative at this time.
Summary of Platelet Activation
Taken together, there is ample evidence for platelet activation in preeclampsia. There is reduced platelet concentration, increased size, reduced lifespan, increased α-granule release, enhanced expression of cell adhesion molecules, and increased thromboxane production. This increased activation, which occurs early in the disease, may either result from an extrinsic factor such as endothelial damage with platelet activation, or it might be intrinsic and antedate pregnancy. Evidence that intrinsic platelet alterations are at least partly responsible comes from findings of platelet binding-site alterations. One example of this is increased platelet angiotensin II binding sites which are apparent in early gestation in some women who subsequently develop preeclampsia. Maki et al. proposed an alternative mechanism caused by diminished circulating platelet-activating factor acetylhydrolase activity. This would lead to decreased platelet-activating factor in normally pregnant compared with nonpregnant women, but not in those with preeclampsia.
Platelet Behavior In Vitro
While the evidence discussed previously points to increased platelet activation in vivo in normal pregnancy, and even more so in preeclamptic women, the data are in contrast to findings from numerous in vitro studies. Specifically, these studies demonstrate increased platelet activation in normal pregnancy compared with nonpregnant women; however, reduced activation is consistently reported in vitro with preeclampsia. In a recent investigation, Burke et al. studied platelet aggregation across normal pregnancy in response to exposure to collagen and arachidonic acid. Platelet aggregation decreased somewhat in early pregnancy, but thereafter it was increased throughout the latter two trimesters.
As indicated, however, others have reported decreased platelet aggregation in women with eclampsia. Horn et al. studied radiolabeled serotonin release in vitro in platelet-rich plasma from preeclamptic patients in response to arachidonic acid. This response was diminished compared with nonhypertensive control subjects. When whole blood is studied in vitro , platelets from women with preeclampsia release less serotonin in response to epinephrine compared with platelets from normally pregnant women. This mirrors the in vitro reduction in platelet aggregation in response to this agonist in preeclampsia. Taken together, these findings suggest that the platelet content of serotonin is reduced in preeclampsia, suggesting that platelets have become activated and released their granule contents prior to in vitro studies.
Platelets, like endothelial cells, contain a constitutive form of nitric oxide synthase, the activity of which was found to be significantly lower in platelets from women with preeclampsia compared with those from normally pregnant women. At the same time, however, platelet nitric oxide and peroxynitrite levels are increased in preeclamptic women compared with normally pregnant controls. Although platelet nitric oxide synthesis may contribute to the vasodilatation of normal pregnancy, it is unclear whether lower activity in preeclampsia results from diminished production, or if it even contributes to its pathogenesis.
Thus, most in vitro studies demonstrate reduced platelet reactivity in preeclampsia compared with normal pregnancy. One possible explanation of the disparate in vivo and in vitro observations is that preeclampsia leads to in vivo activation which results in circulating “exhausted” platelets which are hyporeactive when tested in vitro .
There have been other in vitro studies designed to elucidate mechanisms responsible for increased in vivo platelet activation in preeclampsia. For example, there is evidence that the inhibitory mechanisms that switch off platelet activation responses may be less effective in preeclampsia. Vascular endothelial cell production of prostacyclin, which acts via cyclic adenosine monophosphate (cAMP) to inhibit platelet aggregation, is diminished in preeclampsia compared with normal pregnancy. The resulting tendency to vasoconstriction and platelet aggregation is accentuated by a diminution in platelet sensitivity to prostacyclin. Horn et al. found that there was no alteration in sensitivity to prostacyclin – or other manipulators of cAMP such as thromboxane synthase inhibitors – when platelets from women with gestational hypertension were compared with platelets from normally pregnant women. In reports focusing on preeclampsia, however, pregnancy-induced diminished susceptibility to prostacyclin inhibition was significantly more marked – up to 50% – in preeclampsia. The finding of increased numbers of platelet thromboxane A receptors in women with preeclampsia is also consistent with increased in vivo platelet activation. Arachidonic acid, ADP, and epinephrine all have a thromboxane-dependent component in their mechanism of action. Finally, increased thromboxane receptor density should lead to increases in vivo reactivity.
Expression of Platelet Receptors in Pregnancy and Preeclampsia
The platelet surface is decorated by a plethora of glycoprotein receptors that function to mediate signals from the extracellular milieu to the platelet signaling machinery, often using specific G-protein-coupled receptors as an intermediary. Some receptors, such as GPIIbIIIa, the platelet fibrinogen receptor, undergo transition to an active conformation only upon platelet activation – “inside-out signaling” – enabling the binding of fibrinogen that bridges platelets during platelet activation. In addition to receptors that bind adhesive ligands, platelets display numerous receptors for platelet agonists, such as protease activated receptors, particularly PAR-1, which bind thrombin, P2Y(12) receptors, which bind ADP, and receptors for PGI2, TXA2 and other mediators. The integrated function of these receptors regulates the platelet activation response, and antiplatelet therapies, either approved or in development, target many of these receptors. However, there are very few data available on the expression of these receptors in normal pregnancy or hypertensive pregnancy disorders. A recent study demonstrated that pregnancy-specific glycoproteins, secreted by syncytiotrophoblast, bind GPIIbIIIa, and inhibit fibrinogen binding; however the role of this interaction in regulation of platelet aggregation during pregnancy is uncertain. Polymorphisms in glycoprotein platelet receptors may regulate platelet responsiveness in vitro , and potentially impact thrombosis risk, but their roles in pregnancy outcomes have not been thoroughly studied; a polymorphism in platelet GP6, which mediated collagen signaling, has been recently associated with platelet hyperaggregability and an increased risk of fetal loss in a retrospective analysis, though an increased risk of pregnancy-associated hypertension was not reported. Likewise, a polymorphism of platelet GPIaIIa (integrin α2β1) was also retrospectively associated with early-onset fetal loss, but not hypertensive disease.
Changes in the expression of P2Y(12), PGI2 or TXA2 receptors involved in platelet activation and signaling have not been reported in pregnancy or pregnancy-induced hyptertensive disease, and are a potentially fruitful area of study. One report described decreased affinity of PGI2 receptors in preeclampsia, but this finding has not been reproduced.
Platelet Second Messengers
In an attempt to further elucidate mechanisms underlying in vivo and in vitro changes in platelet behavior, studies have been done to investigate platelet second messenger systems in normal pregnancy and in preeclampsia. The majority of studies have been directed at intracellular free calcium ([Ca 2+ ]) and cyclic AMP. In one study of platelet [Ca 2+ ], Barr et al. found no differences in basal or ADP-stimulated platelet [Ca 2+ ] in either normal pregnancy or preeclampsia. They used the calcium-sensitive indicator quin-2 , which is known to quench increases in platelet [Ca 2+ ] resulting from platelet stimulation. Although they used imprecise methods, they did demonstrate reduced serotonin-stimulated platelet [Ca 2+ ] in preeclampsia compared with normal pregnancy. It had been previously shown that 5-hydroxytryptamine responses were easily suppressed as a result of prior platelet activation.
The calcium-sensitive fluorophore fura-2 has the advantage of weaker calcium-chelating properties than quin-2 . Using this indicator, Kilby et al. found that basal platelet [Ca 2+ ] levels were increased in normal pregnancy and further increased in preeclampsia but not in nonproteinuric gestational hypertension. Whether this increase in platelet [Ca 2+ ] reflects a population of partially activated platelets in preeclampsia, or is a cause of altered platelet reactivity, is unclear. There is some evidence that alteration in stimulated platelet [Ca 2+ ] precedes clinical signs of preeclampsia. In a prospective study of nulliparous women, Zemel et al. found that African-American women who subsequently developed preeclampsia had higher levels of platelet [Ca 2+ ] following stimulation with arginine vasopressin. But while racial differences in platelet [Ca 2+ ] have been reported, Kyle et al. did not find increased vasopressin stimulation of platelet [Ca 2+ ] levels with established preeclampsia in Caucasian women.
Basal cAMP levels do not appear to differ from the nonpregnant state in either normal pregnancy or preeclampsia. Horn et al. demonstrated reduced platelet production of cAMP in pregnancy in response to a range of adenylate cyclase stimulators. They used a sensitive assay based on prelabeling of the metabolic adenine nucleotide pool in platelets with hydrogen 3-adenine. These findings are consistent with the reduction in sensitivity of platelets during pregnancy to inhibition by prostaglandins and other agents which act by increasing levels of cAMP, the inhibitory second messenger. In a cross-sectional study, no differences were found between normal and a heterogeneous group of women including those with nonproteinuric gestational hypertension and preeclampsia. These same investigators did a longitudinal comparative study in a small number of low-risk pregnant women and a group at high risk of developing preeclampsia. They reported that platelets from at-risk women accumulated less cAMP in response to adenylate cyclase stimulators.
Cyclic guanosine monophosphate (cGMP) is another second messenger that inhibits platelet activation. It is synthesized from GTP by the cytosolic-soluble guanylate cyclase enzyme, and in platelets guanylate cyclase is stimulated by nitric oxide, a potent inhibitor of platelet activation. Hardy et al. tested the hypothesis that platelet activation in preeclampsia resulted from underactivity of the inhibitory cGMP system. They found that both platelet cGMP responses to nitric oxide donors and the inhibitory effect of donors on platelet release were increased in platelets from women with preeclampsia compared with normotensive pregnant as well as nonpregnant women. They speculated that upregulation of platelet guanylate cyclase activity may be a compensatory response to impaired nitric oxide production in preeclampsia.
Platelet Angiotensin II-Binding Sites
Specific angiotensin II-binding sites with the characteristics of receptors have been demonstrated on the surface of platelets. Their role is unclear, although it has been suggested that angiotensin II enhances the platelet aggregation response to ADP and epinephrine. Moreover, in vitro studies suggest that angiotensin II increases platelet [Ca 2+ ] levels and this increase is greater in platelets from women with preeclampsia compared with normal pregnancy. Platelets have many of the structural and biochemical characteristics of smooth muscle cells, and similarities between catecholamine-induced changes in both platelet behavior and vascular tone have been described. Measurement of platelet angiotensin II-binding sites has been suggested as an alternative to the labor-intensive and invasive angiotensin II sensitivity test to predict preeclampsia and gestational hypertension. Platelets from normotensive pregnant women exhibit reduced angiotensin II-binding sites, whereas binding site concentrations revert toward nonpregnant levels in preeclampsia. These results mirror the pressor sensitivity of vascular smooth muscle to intravenously infused angiotensin II. This is in contrast to normal pregnancy, in which there is a reduced pressor response to infused angiotensin II, but there is an increased pressor effect prior to the onset of clinical preeclampsia.
The process of coagulation involves a series of enzymatic reactions, which ultimately lead to the conversion of soluble plasma fibrinogen to fibrin clot. Classically, this enzyme sequence is divided into the intrinsic and extrinsic pathways, which both converge in a final common pathway. The major distinction between the two pathways is that the intrinsic pathway is activated from within the bloodstream while the extrinsic pathway begins in the blood vessel walls. In both pathways, clotting factors, numbered I to XIII, are responsible for different interactions and in turn activate the next clotting factor, resulting in a cascading reaction as illustrated in Fig. 17.1 .
The extrinsic pathway is activated at the time of blood vessel injury. Thromboplastin released from damaged cells, together with factor VII and calcium ions, activates factor X, which is the beginning of the common pathway . The intrinsic pathway is initiated by activation of factor XII by collagen. This in turn leads to a series of reactions culminating with the activation of factor X. Activated factor XII also converts prekallikrein to kallikrein, which leads to further activation of XII and activation of the fibrinolytic pathway. Activated factor IX together with factor VIII, calcium ions and platelet factor 3 activate factor X. Activated factor X eventually leads to conversion of prothrombin to thrombin. Thrombin hydrolyzes the peptide bonds of fibrinogen molecules to form fibrin. At the same time thrombin, in the presence of calcium ions, activates factor XIII, which stabilizes the fibrin clot by cross-linking adjacent fibrin molecules. This process is regulated by the fibrinolytic system shown in Fig. 17.1 , composed primarily of plasminogen and regulatory proteins, including antithrombin III, protein C, and protein S.
Coagulation Cascade Factors
In normal pregnancy, the coagulation cascade appears to be in an activated state. Evidence of activation includes increased concentrations of all the clotting factors except factors XI and XIII, with increased levels of high molecular weight fibrinogen complexes. Considering the substantive physiological increase in plasma volume in normal pregnancy such increased concentrations represent a marked increase in production of these procoagulants.
There is considerable evidence that preeclampsia is accompanied by a number of coagulopathic changes when compared with normal pregnant women. Despite this, except for the rare case of preeclampsia complicated by overt clinical disseminated intravascular coagulation, routine coagulation tests are usually normal. This is because clinical tests commonly used are relatively insensitive to minor changes in the coagulation system. There have been reports, however, of covert activation of both the intrinsic and extrinsic coagulation pathways in preeclampsia. Vaziri et al. found changes in plasma coagulation activity of all intrinsic pathway factors in preeclampsia and suggested that such activation was fundamental to the pathogenesis of preeclampsia. The mechanism for such activation is unclear, although endothelial injury and exposure of subendothelial tissue would facilitate activation of factor XII.
One of the most definitive tests of coagulation activity is factor VIII consumption. The test depends on simultaneous measurement of factor VIII clotting activity and factor VIII-related antigen. When the clotting system is activated, circulating levels of both factors increase rapidly as a secondary response, but because factor VIII clotting activity is destroyed by thrombin, its final level is lower than that of the related antigen and the difference between the two is a reflection of factor VIII consumption. During normal pregnancy the levels of factor VIII coagulation activity and factor VIII-related antigen show a proportional rise, thus their ratio remains constant. In preeclampsia, there is an early rise in the factor VIII-related antigen:coagulation activity ratio, which correlates with severity of the disease and the degree of hyperuricemia. While this increased ratio was initially thought to be due to factor VIII consumption, Scholtes et al. found it to be almost entirely due to increased factor VIII-related antigen. This was most marked in cases of preeclampsia associated with fetal growth restriction. Because factor VIII-related antigen is synthesized by endothelial cells and megakaryocytes and is released by aggregating platelets, it is possible that increased levels result from endothelial damage and platelet aggregation rather than increased thrombin action.
During normal pregnancy, plasma fibrinogen concentration substantively increases. Although fibrinogen levels are the same or only slightly increased in women with preeclampsia compared with normal pregnant women, the turnover of radiolabeled fibrinogen is increased in preeclamptic women. Because increased fibrinogen turnover returned to normal with low-dose heparin therapy, it was concluded that it was thrombin-mediated.
The action of thrombin on fibrinogen is a crucial step in the coagulation cascade. Thrombin cleaves two pairs of peptides – fibrinopeptide A and B – from fibrinogen to produce soluble fibrin monomer, which rapidly polymerizes to fibrin. Determination of free fibrinopeptides in blood can be used to measure thrombin activity and fibrinopeptide concentrations are considered to be the best markers of accelerated thrombosis or coagulopathy. Levels of fibrinopeptide A are either elevated or they are unchanged in normal pregnancy; however, most investigators describe increased fibrinopeptide levels in women with preeclampsia compared with normal pregnant women. Moreover, Borok et al. found that the increased serial total fibrinopeptide measurements in preeclampsia correlated with the clinical manifestations of the disease and persisted for 3 to 7 days postpartum.
Regulatory Proteins and Thrombophilia
A number of plasma proteins regulate and maintain the coagulation cascade and its intricate balance between the liquid and solid phases ( Fig. 17.1 ). It seems indisputable that many of the mutations of these factors are associated with a substantively increased risk of venous thromboembolism. Any association between these and adverse pregnancy outcomes is, however, controversial, as recently reviewed by the American College of Obstetricians and Gynecologists.
This glycoprotein is manufactured by the liver. It is an important physiologic inhibitor of coagulation and forms irreversible complexes with all activated factors except VIIIa. There is increased antithrombin III synthesis associated with coagulation activation. Decreased antithrombin III activity indicates increased thrombin binding secondary to increased thrombin generation and is found after thromboembolic events, in disseminated intravascular coagulopathy, and after major surgery.
Antithrombin III activity levels appear to be unchanged in normal pregnancy, although marginal decreases have been described. Decreased antithrombin III activity has been demonstrated in most women with preeclampsia, but not in pregnant women with chronic hypertension. Exacerbations and remissions of the disease were reflected in fluctuations of antithrombin Ill levels, and low antithrombin III concentrations were associated with placental infarctions as well as perinatal and maternal morbidity and mortality. Changes in antithrombin III levels did not, however, correlate with clinical improvement during the puerperium. Antithrombin III activity was reported to begin to decline as much as 13 weeks prior to the development of clinical manifestations in three women who were studied longitudinally and who developed preeclampsia. Diminished antithrombin III activity in preeclampsia is not a consequence of liver dysfunction, but is caused by increased consumption. This led Paternoster et al. to suggest that administration of antithrombin III to women with preeclampsia to normalize the chronic coagulopathy may improve fetal outcome. This was supported by studies in which administration of antithrombin III prevented renal dysfunction and hypertension induced by enhanced intravascular coagulation in pregnant rats. . In a subsequent case-control prospective trial, antithrombin III administration to women with moderate to severe preeclampsia resulted in a significant prolongation of pregnancy, improved neonatal outcomes, and less maternal hemorrhage. Finally, antithrombin III deficiency has been associated with preeclampsia in some but not all. Because of its rarity, it is not possible to ascertain the individual impact of antithrombin deficiency on pregnancy outcome.
This naturally occurring anticoagulant protein serves as a cofactor for activated protein C in the degradation of the activated factors V and VIII by binding to lipid and platelet surfaces. Levels during pregnancy may decrease to those levels found in patients with congenital protein S deficiency. Levels are further diminished in women with preeclampsia compared with normal pregnant women. In one large prospective study of women with severe early-onset preeclampsia, 35% had protein S deficiency. Recent systematic reviews have reported conflicting results regarding the strength of the association between protein S deficiency and preeclampsia. .
This is a serine protease synthesized by the liver and it is activated by contact with thrombin and thrombomodulin on the surface of endothelial cells. It is a potent inhibitor of activated factor V and VIII, and is an activator of fibrinolysis. Protein C is highly sensitive to consumption and reduced levels are found after surgery, thromboembolic events, and in disseminated intravascular coagulation. Protein C levels appear to be unchanged in normal pregnancy compared with the nonpregnant state. Protein C levels have been found to be substantively reduced in preeclampsia compared with normal pregnancy.
Activated Protein C Resistance
This mutation results in a limitation in anticoagulant response to activated protein C which predisposes to venous thromboembolism. The cause was identified as a point mutation in the factor V gene at nucleotide position 1691, resulting in an arginine-to-glutamine substitution. This reduces the sensitivity of the factor V protein to inactivation by activated protein C – hence, activated protein C resistance – resulting in a procoagulant state and an increased risk of thrombosis. The resultant mutation is termed factor V Leiden because the research was carried out in this Dutch town. The trait is inherited in an autosomal dominant manner with the risk of thrombosis increased seven-fold in heterozygotes and 80-fold in homozygotes. Studies have shown that the distribution of the factor V Leiden mutation varies in different populations – it is found in about 5% of Caucasians, Europeans, Jews, Arabs, and Indians while it is virtually absent in Africans and Asians.
Since the discovery and characterization of activated protein C resistance, a myriad of largely retrospective studies have examined the association between factor V Leiden mutation and a range of adverse pregnancy outcomes including preeclampsia. One metaanalysis of many of these studies suggested that factor V Leiden is associated with a 2.04-fold (95% CI 1.23–3.36) increased risk of severe preeclampsia. Factor V Leiden accounts for approximately 70% of women with activated protein C resistance; the remainder are factor V Leiden negative (APCR FVL– ) and they are considered to have acquired activated protein C resistance. The effect of APCR FVL– on pregnancy outcomes is not well characterized. Lindqvist et al. perfomed a prospective study of 2480 unselected women in early pregnancy. The APCR FVL– cohort of 106 women did not have an increased risk of preeclampsia. Similar findings from a recent Australian study have also been reported. According to the American College of Obstetricians and Gynecologists, data are inconclusive that factor V Leiden mutations increase the risk of any adverse pregnancy outcome except for venous thromboembolism.
This procoagulant protein is the inactive precursor of thrombin. It results from a G-to-A mutation at nucleotide 20210 in the prothrombin gene that causes excessive production and increased circulating prothrombin levels. Like the factor V Leiden variant, this mutation shows significant ethnic variation, being most prevalent in southern European populations, with an overall population prevalence of 2–3%. Controversy remains as to whether it is associated with preeclampsia. In a number of case-control studies there does not appear to be a significant association. In some small studies, however, the prothrombin mutation does appear to be associated with fetal-growth restriction and placental abruption.
These autoimmune antibodies include lupus anticoagulant, anticardiolipin antibodies, and anti-β 2 -glycoprotein. Amongst other adverse effects, these autoantibodies are associated with recurrent pregnancy loss, preeclampsia, and fetal-growth restriction. They are also associated with a four-fold increase of thrombosis or thrombocytopenia. The presence of antiphospholipid antibodies – anticardiolipin antibodies or lupus anticoagulant or both – significantly increases the risk of developing preeclampsia (RR 9.72, 94% CI 4.34–21.75). There is evidence that treatment with aspirin and low-molecular-weight heparin (LMWH) can ameliorate these effects. While large interventional trials have studied the effect of aspirin on the recurrence rate of preeclampsia, only limited data are available regarding LMWH. In a small study from the Netherlands, the recurrence rate of preeclampsia was similar in 26 women with thrombophilia following treatment with LMWH and aspirin and 32 women receiving aspirin alone. Several studies since, however, have demonstrated improved outcomes with a combination of LMWH and aspirin when compared with aspirin alone in women with a history of preeclampsia in a previous pregnancy. Larger and appropriately powered randomized trials are needed to clarify this issue.
The end-product of the coagulation cascade is fibrin formation. The main function of the fibrinolytic system is to remove excess fibrin deposited as a consequence of thrombin activity (see Fig. 17.1 ). Plasmin, which is the main protease enzyme in this system, originates from plasminogen secreted by the liver. The activation of plasminogen into plasmin is through plasminogen activators which are serine proteases. These include tissue- and urokinase-plasminogen activators – t-PA and u-PA – and they are secreted by endothelial cells, monocytes, and macrophages. Both act upon plasminogen to convert it to plasmin and in so doing trigger a proteolysis cascade that causes thrombolysis. Plasmin cleaves and converts t-PA and u-PA into two-chain proteases, which exhibit higher proteolytic activity, implying a positive feedback for the fibrinolytic cascade.
A negative feedback is therefore essential for the fibrinolytic pathway and thus the activity of plasminogen activators is balanced by plasminogen activator inhibitors. Plasminogen activator inhibitor type 1 (PAI-1) is a serine protease inhibitor that functions as the principal inhibitor of t-PA and u-PA. The other PAI, plasminogen activator inhibitor-2 (PAI-2), is secreted by the placenta and is only present in significant amounts during pregnancy. In addition, plasminogen activator inhibitor-3 (the protease nexin) acts as an inhibitor of t-PA and urokinase.
There are other fibrinolytic inhibitors. A2-antiplasmin (α2-AP) is a single-chain glycoprotein that reacts with plasmin to form a plasmin–α2-AP complex which is incapable of breaking down fibrin. Thrombin activatable fibrinolytic inhibitor – TAFI – is a glycoprotein synthesized by the liver and also found in platelet granules and it acts as an enzyme that may modulate fibrinolytic activity. Finally, α 2 -macroglobulin – α2-M – is synthesized mainly by the liver and is a general inhibitor of both coagulation and fibrinolysis, acting as a scavenger. In the fibrinolytic system, α2-M inhibits the action of plasmin and kallikrein, while in coagulation it inhibits thrombin.
Fibrinolysis in Normal Pregnancy
Studies of the fibrinolytic system in pregnancy have produced conflicting results. The majority of early studies suggested that fibrinolytic activity is reduced in normal pregnancy, reaching a nadir in the third trimester.
Conversely, investigators have reported that plasma plasminogen concentrations are increased in normal pregnancy and more recent studies have shown that t-PA and u-PA levels increase in pregnancy, suggesting an activation of the fibrinolytic system. To counter such activation, there is a several-fold increase in PAI-1 levels and placental production of PAI-2. A progressive increase in serum fibrin degradation products and D-dimers has also been observed throughout pregnancy. Since D-dimers reflect both fibrin polymerization and breakdown, fibrinolysis has been considered active during pregnancy.
Fibrinolysis in Preeclampsia
The fibrinolytic pathway in preeclampsia has been the focus of intense investigation. Overall, increased levels of circulating t-PA levels have been described in preeclamptic compared with normal pregnancies. Most, but not all, studies report increased levels of PAI-1 in preeclamptic pregnancies. By way of contrast, however, other studies have reported a significant reduction or no difference in PAI-1 plasma levels in preeclampsia. Because both t-PA and PAI-1 are synthesized by endothelial cells, their increased levels may reflect endothelial dysfunction. In contrast, PAI-2 is produced by the placenta and has been reported as being significantly decreased in severe preeclampsia, possibly reflecting placental insufficiency. In support of this, reduced levels of PAI-2 have also been reported in pregnancies complicated by fetal-growth restriction, with or without concomitant preeclampsia.
The discrepant results for PAI-1 might in part be explained by the fact that preeclampsia may be a heterogeneous disorder with two principal phenotypes often characterized by the time of presentation (see Chapter 6 ). Early-onset disease presents between 24 and 34 weeks gestation and late-onset disease presents closer to term. Early-onset disease is thought to be predominantly driven by uteroplacental insufficiency, which is less obvious in women with term preeclampsia. It is possible that increased t-PA and PAI-1 antigen levels found in preeclampsia could be regarded as markers for a phenotype which presents predominantly with endothelial dysfunction, while reduced PAI-2 could reflect a uteroplacental insufficiency-driven phenotype. In support of this, Wikstrom et al. demonstrated decreased PAI-2 levels, increased placental oxidative stress, and increased PAI-1:PAI-2 ratio in early-onset, but not in late-onset preeclampsia.
Studies of fibrin–fibrinogen degradation products in preeclamptic women have been conflicting. These have included reports of D-dimer levels, a well-established clinical laboratory marker of fibrin polymerization and breakdown in vivo (see Fig. 17.1 ). Most, but not all, studies have shown increased dimer levels in women with preeclampsia compared with those in normal pregnancy. A recent metaanalysis confirmed this association with third-trimester preeclampsia, but the authors highlighted the need for longitudinal studies throughout pregnancy in order to fully elucidate any prognostic value.
At this juncture, it is fair to conclude that assessment of the fibrinolytic system in preeclampsia is difficult to interpret. There have been reports of reduced fibrinogen levels, increases in t-PA but not u-PA, increases in PAI-1 but not PAI-2, and equivocal evidence of increased fibrin–fibrinogen degradation products. It is difficult to establish whether any of these changes contribute to or merely reflect changes induced by preeclampsia. Altered endothelial cell function in preeclampsia may result in increased release of both tPA and PAI-1 and reduced levels of PAI-2 might reflect either impaired placental function or a disorder in the fibrinolytic mechanism.
A polymorphism in the plasminogen activator inhibitor-1 gene has been described and has been examined in association with obstetrical complications. Homozygosity for the 4G insertion in the promoter region results in higher circulating levels of PAI-1 and is associated with a higher risk of thrombosis. Glueck et al. found homozygosity for the 4G allele to be twice as common in women with obstetrical complications including preeclampsia compared with normal pregnant control women. It was also seen more frequently in association with other recognized thrombophilias in women with complications. Two recent metaanalyses have confirmed that this polymorphism is a likely susceptibility variant for preeclampsia.