The cost-benefit ratio of screening pregnant women for thrombophilia

Introduction

The Norwegian researcher Egeberg coined the term ‘thrombophilia’ in 1965, after having described for the first time the association between venous thromboembolism, occurring in the absence of any specific causes, and an inherited antithrombin III defect¹. The term thrombophilia is now used to define all those inherited, acquired and/or transitory conditions that predispose to the onset of arterial or venous thromboses^2,3. The crucial point regarding research into thrombophilia is the evaluation of the correct functioning of haemostasis. Haemostasis is currently defined in terms of ‘haemostatic balance’, a terminology that envelops a vast series of physiological mechanisms aimed at maintaining the blood fluid while it is in the vascular system, but preventing excessive haemorrhagic loss following endothelial damage or breaks in the wall of the blood vessel. A series of events is activated as a result of a lesion in a blood vessel. These events can be divided into successive phases according to a cascade model: the blood vessel-platelet activation phase (primary haemostasis), coagulation phase (secondary haemostasis), clot dissolution (fibrinolysis), and repair of the endothelial damage. The particular characteristics of the haemostatic process are its localisation, amplification and modulation that, in physiological conditions, are in perfect dynamic equilibrium⁴. These complex haemostatic functions are the result of the integrated and finely regulated action of many components, of which the main ones are the vascular endothelium, platelets, circulating and transmembrane proteins and calcium ions. The activation and amplification of the cascade and stabilisation of the coagulum are regulated by the balance between the activity of specific proteases and that of their respective allosteric and enzymatic inhibitors⁵. Abnormalities at any point during the coagulation cascade can cause pathological changes, characterised by haemorrhages or intravascular thromboses, depending on whether there is an insufficient or excessive response to endothelial damage⁶.

In 1856, the Prussian pathologist Rudolf Virchow first proposed his hypothesis, which was subsequently shown to be correct and just as relevant nowadays, to explain the pathogenesis of thrombosis⁷. He suggested that three factors were sufficient and necessary to produce thrombosis: (i) hypercoagulability, (ii) stasis and, (iii) endothelial damage. These factors, present to variable degrees in the pathogenesis of venous thrombosis, also concur to increase the risk of thromboembolism during a normal pregnancy. Indeed, a state of hypercoagulability develops during pregnancy; this is caused by both increases of certain procoagulant factors and decreased effectiveness of inhibitory systems. Furthermore, the mechanical obstruction caused by the foetus and the vasodilatory effects mediated by the altered oestrogen/progesterone ratio result in increased blood pressure and stasis in the lower limbs. Together, these phenomena are responsible for the endothelial damage and compromise primary and secondary haemostasis.

This review describes the main haemostatic changes occurring during pregnancy, focusing on aetiopathological aspects, the diagnostic procedures for evaluating the risk of thrombophilia and the appropriate diagnostic measures.

Alterations of haemostasis during pregnancy

During pregnancy, a series of changes occur; these changes involve all the haemostatic balance effectors, which are traditionally divided into procoagulant, anticoagulant and fibrinolytic factors (Table I).

Table I.

Coagulation factor changes in pregnancy

	No change	Increase	Decrease
Procoagulants		I,V,VII, VIII, IX, X	XI
Anticoagulants	Protein C	Soluble TM	Protein S
Adhesion proteins		VWF
Fibrinolytic proteins	TAFI	PAI-1, PAI-2	t-PA
Antiphospholipid antibodies	APLA	Micro-particles
Placental variations		T F	TFPI

TM, Trombo-Modulin; TAFI, Thrombin Activator Fibrinolysis Inhibitor; PAI-1 and PAI-2, Plasminogen Activator Inhibitor type 1 and 2; t-PA, tissue Plasminogen Activator; APLA, Anti PhosphoLipid Antibodies; TF Tissue Factor; TFPI, Tissue Factor Pathway Inhibitors.

Changes in procoagulant factors

The data on variations in procoagulant factor levels are sometimes contradictory, in some cases because of different methods used and the characteristics of the populations under study. The most likely cause of the conflicting results is that some studies are outdated; thus, re-evaluations are necessary using more modern and standardised equipment and methods.

The concentrations of various coagulation factors (factors V, VII, VIII, IX, X, and XII) and von Willebrand factor (VWF) increase significantly in pregnancy, accompanied by a pronounced increase in fibrinogen, up to twice the normal reference values^8–15.

The concentration of factor VIII (FVIII) increases progressively during the first period of pregnancy, as does that of VWF, so the FVIII/VWF ratio remains constant^16–18. In the third trimester there is, however, a change in the FVIII/VWF ratio, which is due to the reduction of FVIII in the last few weeks of pregnancy¹⁶.

During the third trimester there is also a considerable increase in factor VII (FVII) and its activated fraction (FVIIa)^16,19–21. Any increase of factor IX (FIX) is usually small and has been described only occasionally²², while the changes in the concentrations of factors XI (FXI) and XII (FXII) are more widely debated.

Some Authors^22,23 have reported a significant reduction of FXI, while others have described that the levels of this clotting factor remain substantially stable²⁴. The concentration of FXII has also been described to remain fairly stable throughout pregnancy²⁴, but there are reports of increases in the third trimester²⁵. The concentrations of factors II (FII) and X (FX) appear to substantially unaltered or slightly increased^16,24,25. Factor XIII (FXIII), or fibrinstabilising factor, appears to diminish²⁶, while tissue factor (TF) remains constant, despite it being less expressed by monocytes^27,28.

Changes is physiological inhibitors of coagulation

Coagulation is regulated by various physiological inhibitors; normally these have a greater effect than procoagulant factors, thus preventing the formation of thrombi within the undamaged vascular system. It has been found that there are changes in both the activity and concentration of plasma coagulation inhibitors during pregnancy, unbalancing the overall system towards a procoagulant state. The most studied physiological inhibitors are protein C, protein S, antithrombin, thrombomodulin and endothelial protein C receptor (EPCR).

Protein C, a vitamin K-dependent factor, circulates as a pro-enzyme and is activated by the thrombin-thrombomodulin complex. It acts by inhibiting FV and FVIII, co-factors of the tenase and prothrombinase complexes, respectively^29,30. The protein C system involves numerous factors in various different reactions, including: a) factors involved in the activation of protein C through binding with thrombin-thrombomodulin, b) co-factors that regulate the proteolytic activity of protein C, and c) serine proteases that inhibit protein C itself ^31–36. The average half-life of activated protein C is approximately 20 minutes³⁷ and is strictly dependent on a series of inhibitors such as α₁–antitrypsin, α₂ –macroglobulin and protein C inhibitor (PCI) that are involved in its inactivation and degradation^31,33. Protein C has recently been described to have anti-inflammatory and anti-apoptotic actions, which contributed towards clarifying a presumed interaction between the coagulative and anti-inflammatory systems^30,38,39. During pregnancy, the levels of protein C are stable or increase slightly; an increase in activity is observed only in the post-partum period⁴⁰. In more detail, it seems that the levels of protein C increase during the second trimester and decrease during the third, to then return to normal by about 5 weeks after the birth. However, despite these variations, the levels remain within normal limits^12,41.

Following the description of activated protein C resistance (APCR) by Dahlback et al. in 1993⁴² and the identification of the genetic mutation of FV responsible for this resistance⁴³, numerous clinical studies have highlighted the importance of APCR in the aetiopathogenesis of deep vein thrombosis in pregnancy. Conversely, many other studies have evaluated the influence of pregnancy on APCR. Although the results are quite heterogeneous, depending on the methods used to determine APCR, the period of pregnancy in which the tests were carried out and the characteristics of the pregnant women, they do show that pregnancy has a substantial influence on APCR^44–51. APCR was initially associated only with the presence of the FV Leiden mutation, but APCR secondary to antiphospholipid antibodies or neoplasms has since been described^52,53. As reported in numerous scientific papers, hormone replacement therapy with oestrogens and progesterones also causes a variation in APCR and many studies have been conducted to try to clarify this supposed influence^54–57.

Protein S is another vitamin K-dependent protein; 60–70% of it circulates in the blood bound to a regulatory protein of complement, the C4b-binding protein (C4bBP)⁵⁸. Only the free form of protein S acts as a co-factor for activated protein C^59–61. It now appears certain that there is a progressive reduction of both overall protein S levels and free protein S during pregnancy^12,62–65.

Thrombomodulin is a membrane glycoprotein expressed mainly by endothelial cells and trophoblasts involved in thrombin and protein C regulation^66–68. Endothelial cell protein C receptor (EPCR) aids the activation of protein C by the thrombin-thrombomodulin complex on the surface of endothelial cells⁶⁹. The soluble form of thrombomodulin can be found in the plasma and urine and can be used as a marker of endothelial damage. Thrombomodulin levels appear to be constantly elevated during pregnancy, as a result of neutrophil activation which increases proteolysis, and normalise in the post-partum period^70,71.

Antithrombin levels are higher in women than in men, increase with age in women and are reduced in association with the use of oestrogen-progesterone^12,72. There is no reliable evidence of a reduction of antithrombin levels during pregnancy, apart from in some particular cases at the end of the pregnancy⁷³. The levels of antithrombin do, therefore, remain substantially stable during pregnancy, although some Authors have observed slight reductions towards the end of pregnancy, without these being associated with any particular thrombotic complications⁷².

Changes in the fibrinolytic system

The activity of the fibrinolytic system is diminished during pregnancy but returns to normal within 1 hour after birth. D-dimer levels are notably increased during labour and Caesarean section, particularly as a result of surgical trauma. Despite this, it is well known that pregnancy causes a progressive increase in plasma D-dimer levels, which are strictly related to gestational age⁷⁴. In contrast, tissue plasminogen activator (t-PA) activity diminishes during pregnancy⁷⁵ not only because of an increase in plasminogen activator inhibitor-1 (PAI-1), but, above all, because of an increase in plasminogen activator inhibitor-2 (PAI-2)^75–78. PAI-1 and PAI-2 levels increase during pregnancy, their activity being five times higher at 35 weeks than at 12 weeks, to then normalise by 5 weeks after the delivery⁷².

The main causes of thrombophilia

Despite the fact that in the last few years new scientific discoveries have clarified some of the triggering events and predisposing factors responsible for thromboembolism, in 50% of cases the causes remain unknown. In the last 10 years, much research has been carried out to investigate inherited, acquired and/or transitory causes of venous thromboembolism in pregnancy, to the point that evaluating thrombophilia in pregnancy has assumed a crucial role not only in the diagnosis but also in the prophylaxis and treatment of thrombotic episodes.

The causes of thrombophilia have also gained a central place in the evaluation of obstetric complications of a vascular nature that alter tissue perfusion causing foetal distress. The main thrombophilic factors can be divided into inherited and acquired types^79–81.

As far as concerns inherited factors, to which most research has been dedicated, the FV Leiden mutation^81–83, the G20210A polymorphism of the prothrombin gene^84,85, the C677 mutation of methylenetetrahydrofolate reductase (MTHFR)^86–87, and deficiencies of protein C, protein S and antithrombin^88,89, all play important roles. More widely discussed inherited defects are polymorphisms of the genes for PAI-1 (4G/4G)^90.91, FXIII^92–94 and FVIIa⁹⁵. It has also been hypothesised that polymorphisms of the thrombomodulin gene are associated with onset of deep vein thrombosis^96,97, although other larger studies have contradicted this hypothesis^98.99, making it necessary to reflect on the research and routine evaluation of polymorphisms for which scientific data are not unequivocal. Polymorphisms of the APCR gene have also been evaluated; the preliminary data have not been supported by clinical findings^100.101. Other thrombophilic conditions have a mixed aetiology, including elevated levels of FVIII^102,103, FIX¹⁰⁴, FXI^105,106, APCR not attributable to the FV Leiden mutation¹⁰⁷, hyperhomocysteinaemia¹⁰⁸, elevated levels of thrombin activatable fibrinolysis inhibitor (TAFI)¹⁰⁹, dysfibrinogenaemia¹¹⁰ and hyperfibrinogenaemia¹¹¹.

Acquired causes of thrombophilia include age^79,112, the presence of antiphospholipid antibodies^113,114, cancer^115–117, and prolonged breastfeeding⁷⁹. Orthopaedic surgery, in particular hip replacements and knee operations, is an important risk factor^118,119, as is major trauma^120,121. There is a very large body of data regarding the association between oral contraceptives and thrombosis, which appear to be a further, but no less important, acquired risk factor in women of fertile age^56,122. Pregnancy itself is also a risk factor for the onset of venous thrombosis because of various anatomical (venous stasis as a consequence of foetal pressure on the iliac veins and inferior vena cava) and biochemical (alteration in the concentration and function of the coagulation factors) conditions. These conditions are of considerable importance in the evaluation and management of pregnant patients, particularly those with other, above described, acquired causes of thrombophilia^{51,80,123–128}.

Incidence and prevalence of inherited thrombophilia in pregnancy

The risk of deep vein thrombosis (DVT) is six times higher in pregnant women than in non-pregnant women¹²⁷, with an incidence that varies between 0.6/1,000 pregnancies for women less than 35 years old to 1.2/1,000 pregnancies for those older than 35¹²⁹. The incidence of DVT in the post-partum period also varies with age, ranging between 0.3/1,000 for women less than 35 years old to 0.7/1,000 for women over 35 years old¹²⁵. The frequency of DVT does not vary throughout the whole of the pregnancy, although it appears to be drastically increased in the post-partum period. It is worth noting that the incidence of thrombosis differs according to the type of delivery, with the incidence being between 0.1–1.2% for vaginal deliveries and up to 2.2–3% following Caesarean sections^123,129. The heterogeneity in the reported incidences of DVT in pregnancy is explained by the fact that only in the last 15–20 years have reliable instrumental and laboratory methods become available to confirm the diagnosis of DVT.

The most frequent obstetric complications that can be traced back to vaso-occlusive alterations in placental perfusion are recurrent foetal loss, intra-uterine growth retardation (IUGR) and pre-eclampsia. Pulmonary embolism remains the main cause of maternal mortality in the western world, although the global incidence of this complication has decreased considerably in the course of the last few years^123,126,130.

In brief, pregnancy appears to be an independent risk factor¹²⁹ for the onset of DVT and the risk increases considerably in the concomitant presence of inherited or acquired thrombophilic conditions. Despite this association between thrombophilia, DVT, and pregnancy, the aetiopathogenesis of more than 50% of the cases of venous thrombosis remains unknown, which suggests the presence of other factors, whether environmental or related to the physiopathology of haemostasis, which have not yet been clarified. Thrombophilic disorders significantly increase the risk of venous thrombosis in pregnancy and are predisposing factors for vascular complications. The prevalence of antithrombin deficiency in the general population varies from 0.02–0.17%, with an incidence of 1.1% in subjects with DVT. Of the inherited forms of thrombophilia, antithrombin deficiency is the most serious, as witnessed by the fact that more than 50% of subjects with this deficiency develop thromboembolic events in their lifetime^88,89. The prevalence of dysfunction of the protein C and protein S systems is between 0.14–0.5% in the general population and 3.2% in subjects with venous thromboses. The risk of thrombosis in pregnancy varies from 3–10% for patients with protein C deficiency, 0–6% for patients with protein S deficiency and much lower for those with antithrombin deficiency. During the post-partum period, the risk of thrombosis rises to 7–19% among women with protein C deficiency and to 7–22% among those with protein S deficiency¹²⁴. It should be noted that protein S levels vary considerably in pregnancy; protein S activity below the reference range is found in 25% of pregnant women in the first trimester, in 60% in the second trimester and in 83–100% in the third trimester⁴⁰.

APCR is present in 3–7% of the general white population and in 20–30% of patients with venous thromboembolism;⁸⁰ in most cases the APCR is attributed to the FV Leiden mutation. APCR was found in 78% of women undergoing investigations for potential thrombophilia and FV Leiden in 46% of the cases^131,132.

The polymorphism of the prothrombin gene promoter (G20210A) is sometimes associated with elevated levels of FII. This substitution is present in about 2–3% of the Caucasian population, while it is rarer in Africans, Asians and native Americans, and is associated with a 3-fold increased risk of having a thromboembolic event⁴. In a recent study, the incidence of DVT in pregnancy was 0.67/1,000. The probability of developing thrombosis in pregnancy was 0.03% for pregnant women without mutations, 0.25% in those with FV Leiden, 0.5% in FII polymorphism carriers and 0.1% in patients with protein C deficiency. The combined probability of thrombosis in subjects with both FV Leiden and FII polymorphism reached 5%¹³³. Hyperhomocysteinaemia is often associated with homozygous mutation of the MTHFR gene (C677T)⁸¹, which has a prevalence in the general population of about 8–10% (Table II)^86,133.

Table II.

Prevalence of the principal thrombophilic disorders in pregnancy

Disorder	Defect	% in general population	% in patients with first DVT	Odds ratio
Antithrombin deficiency	Reduced AT III levels	0.07	1	10–20%
Protein C deficiency	Reduced protein C levels	0.3	3	6–8%
Protein S deficiency	Reduced protein S levels	0.2	3	2–6%
Heterozygous FV Leiden	G1691A mutation of factor V	5–8	20	4–8%
Homozygous FV Leiden	G1691A mutation of factor V	0.06	1.5	80%
Prothrombin gene mutation	G20210A mutation of factor II	3	6	2–4%
Hyperhomocysteinaemia	Elevated homocysteine	5	10	2–3%
Homozygous MTHFR C677T	MTHFR C677T gene mutation	10–20	11–12	0.7–2%
Antiphospholipid antibodies	Antibodies present	2	10–15	9%
Activated protein C resistance (Leiden correlation excluded)	Factors interfering with protein C activity	8–11	24	2–4%

DVT, deep vein thrombosis; AT, antithrombin.

Screening for thrombophilia

In an overall assessment of the value of thrombophilia screening in young women wanting to start a family, it seems particularly important to establish a diagnostic-therapeutic protocol suitable for evaluating: firstly, the candidates to whom the screening should be applied; secondly, the most appropriate analytical methods, and, lastly, the interpretation of the results.

There appears to be a substantial agreement in the scientific literature that the subjects who could benefit from screening are those for whom diagnostic and therapeutic information of clinical relevance could be obtained. Testing for thrombophilia would seem to be justified in a limited number of patients who have precise clinical characteristics and whose risk of vascular complications during pregnancy appears to be substantially raised. Systematic screening of all pregnant women is not feasible, as confirmed by a series of clinical evaluations of this issue^134–140. In this specific circumstance, it is also misleading to extend the concept of screening, since, from a legislative point of view, this term means a service supplied by the national health service for a specific category of people, with the purpose of identifying those at risk of developing a specific pathology or preventing its complications by using a simple, inexpensive test.

In the light of all this, there is still no consensus on the cost-benefit ratio of this approach¹⁴¹. In depth investigations for thrombophilia appear to be justified in selected individuals, who are known to have risk factors or have an obstetric history suggestive of such factors, because of complications probably caused by haemodynamic alterations. The patient’s clinical history must, therefore, identify obstetric complications, such as repeated abortions, IUGR, pre-eclampsia and placental abruption. Only the presence of a history suggestive of thrombophilia can be accepted as an essential preliminary to justify carrying out subsequent in depth diagnostic tests. Indeed, looking for FV Leiden is not only unadvisable in terms of the cost-benefit ratio¹⁴², but could also provoke unjustified anxiety and fear. A series of features concur to identify patients who could benefit from further diagnostic investigations. As already said, evaluation of the patient’s clinical and family history is crucial^143,144. A personal or family history positive for complications in pregnancy can suggest the need to carry out in depth diagnostic investigations, since the most common thrombophilic conditions are frequently associated with obstetric complications.

Abortions and recurrent foetal loss

From 1–3% of fertile women suffer recurrent abortions. Although in the majority of cases there is an identifiable cause, some potential aetiopathogenetic factors have been described. These include anatomical changes, chromosomal alterations, metabolic and endocrinological imbalances and autoimmune disorders^143,144. In the last 20 years, there has been increasing emphasis on the role of inherited and acquired thrombophilia in patients with a history of recurrent abortions or complications due to vascular malfunction. Different classifications have been used, depending on the period of gestation in which the complication occurs. Thrombophilic changes are found in about 49–65% of women with vascular complications during pregnancy, but in only 18–22% of women with normal pregnancies, indicating a 3 to 8-fold relative risk^82,145. Since an association between heterozygosity for FV Leiden and unexplained recurrent foetal loss has been observed in more than 30% of cases, it is estimated that the relative risk of gestational complications caused by this mutation is between 2 and 5. This conclusion has, however, been challenged by the results of other studies, although these studies have certain objective limitations, such as the small number of people selected and having included patients with first-trimester miscarriages, which, as known, are usually attributed to the most serious causes of thrombophilia^146–148. In the last few years^149–153, research carried out using more selected populations has shown that the risk of recurrent abortions in female carriers of the FV Leiden mutation is double that in women without the mutation; furthermore, the risk in pregnant women homozygous for the mutation appears to be double that in women who are heterozygous for the same mutation. A recent meta-analysis of more than 3,000 patients confirmed the association between FV Leiden and recurrent abortions, quantifying the relative risks in both the first trimester (2.1) and in the successive ones (7.8)¹⁵³.

APCR is an independent risk factor with respect to the presence of FV Leiden for thrombosis^154,155 and it has been suggested that it might increase the predisposition to various thrombotic complications in pregnancy. Indeed, APCR is found in between 9–38% of women who have an unexplained pregnancy loss, but in less than 3% of pregnancies without complications^156,157. These data suggest that in pregnancy APCR acts independently of the FV Leiden mutation in increasing the risk of recurrent abortions and other vascular complications¹⁵³. The above-mentioned polymorphism of the prothrombin gene is associated with increasing prothrombin activity, thus causing a 2 to 4-fold increased risk of developing a deep vein thrombosis⁸⁴. The mutation has been observed in many gynaecological studies of patients who have had a spontaneous abortion. In particular, the mutation is present in 4–9% of women who have had recurrent abortions (the majority in the first trimester), with respect to a prevalence of 1–2% in uncomplicated pregnancies, increasing the overall risk to 2–9^158,159.

Deficiencies of natural anticoagulants, such as protein C, protein S and antithrombin, are very rare and, overall, occur in fewer than 2% of the general population. These deficiencies seem to increase the probability of gestational complications, even if not all studies are concordant in attributing a tangible risk^160–163. The discordance in published results is considered to be due to the low prevalence of the deficiencies in the general population and to the small number of cases studied. Protein C deficiency is associated with a 2–fold increased risk of foetal loss both in the first trimester and at the end of the pregnancy^153–163. Protein S deficiency is associated with a more consistent risk (from 3–40 times higher)^160–164, since the prevalence of this deficiency in pregnant women with complications is 5–8%, compared to less than 0.2% in pregnant women without complications. Antithrombin deficiency is also associated with an increased risk (from 2 to 5-fold) of recurrent foetal loss. There are numerous studies on this issue and their results and conclusions are sometimes in stark contrast with each another. Despite all this, some meta-analyses seem to indicate that pregnant women with antithrombin deficiency do have an increased risk of multiple abortions. Hyperhomocysteinaemia is an independent risk factor for the development of venous thromboembolism^165–167.

Some clinical studies have confirmed an increased prevalence of hyperhomocysteinaemia in women who have had spontaneous or recurrent abortions^168–171. The interpretation of these results appears to be rather problematic given the difficulty in defining reliable reference values for plasma homocysteine, when to carry out the analysis, and the method to use as well as the effect of possible vitamin supplementation, particularly with folates and B12, which notoriously have an influence on the metabolism of this sulphated amino acid. Hyperhomocysteinaemia is observed in 17–27% of women after a first or recurrent miscarriage, compared to in 5–16% of a control population, increasing the relative risk, from 3–7^167,169. A study based on a population of more than 5,000 pregnant women showed that the risk of abortion was double among those women with homocysteine levels above 10 mmol/L. Histopatholological examination of tissue samples from women who suffered a pregnancy loss in the first trimester seemed to show a causal link with the hyperhomocysteinaemia, mediated by the vascular changes of the chorionic villi that would predispose to the abortion. A point mutation of the MTHFR gene produces a ‘thermolabile’ form of the enzyme which has reduced enzymatic activity. Homozygosity for this mutation is very common in the general population (10–20%) and is a predisposing factor for hyperhomocysteinaemia. Heterozygosity for this mutation is more common (the prevalence being approximately 46%), but its association with hyperhomocysteinaemia is much debated, given that this condition, depends largely on the amount of vitamin supplementation provided in the diet or by specific drugs (folin)¹⁷² Women with dysfibrinogenaemia have a higher incidence of developing thrombosis and recurrent foetal loss in the first trimester of pregnancy¹⁷³.

At present, many situations are being investigated to evaluate the risk of thrombophilia in women with recurrent foetal loss. Examples are recurrent abortions or foetal death after the tenth week of gestation, complications that are explicitly included among the clinical criteria for defining antiphospholipid antibody syndrome proposed by the Consensus Conference held in Sapporo in 1999¹⁷⁴. Other thrombophilic conditions, frequently the subject of research, are thrombomodulin and protein C receptor mutations, PAI-1 gene polymorphisms (4G/4G), FXII or protein Z deficiencies and increased levels of circulating procoagulant micro-particles^175–179. In the evaluation of thrombophilia in recurrent abortions, particular attention should be given to the association of several risk factors, since their effect on the relative risk may be multiplicative and not simply additive. For example, the incidence, of heterozygosity for FV Leiden and the FII point mutation is approximately 1 in 1,000.

Intra-uterine growth retardation

The observations on IUGR are more limited, and, for certain aspects, more contradictory. Thrombophilic defects are found in approximately 60–70% of women with IUGR, whereas the incidence in the general population does not exceed 18%, with a 4-fold estimated relative risk¹⁸⁰. FV Leiden and the polymorphism of the prothrombin gene have been observed in, respectively, 8–35% and 7–15% of pregnant women with IUGR, compared to in 2–4% of women with a normal pregnancy. Both mutations increase the risk of IUGR, by 7–13 times for FV Leiden and by 4–9 times for the polymorphism of the prothrombin gene^181,182. Although some studies suggest that APCR not due to FV Leiden and hyperhomocysteinaemia are strictly associated with IUGR and low birth weight^183,184, other studies have not confirmed any association between thrombophilic disorders and IUGR^172,185.

Pre-eclampsia and placental abruption

According to reliable estimates, pre-eclampsia complicates up to 3% of pregnancies^186,187. Thromboses of the spiral arteries of the placenta produce tissue hypoperfusion with endothelial dysfunction and changes in the micro-circulation. Although numerous studies have shown that a large series of thrombophilic disorders are associated with other gestational problems, there is not currently conclusive evidence on the role of thrombophilic changes in the aetiopathogenesis of pre-eclampsia and placental abruption. This may be a consequence of the changing definitions of these complications, the ethnic groups studied, the selection criteria and the size of the population. Some retrospective studies have shown the presence of at least one cause of thrombophilia in 40–72% of women with pre-eclampsia, with respect to a prevalence of 8–20% in women without complications^188–192. In particular, thrombophilic disorders were found in 67% of women with severe pre-eclampsia, but in 20% of women with normal pregnancies, which translates into a relative risk of 8¹⁸⁹. Furthermore, it should be stated that other studies have not shown any significant association between thrombophilia and pre-eclampsia^193–196. Despite this, it is undisputable that women with a positive history for thrombophilia have a 3-fold higher risk of developing thromboses¹⁹⁷.

Identifying patients for thrombophilia screening remains a crucial matter, principally in order to be able to define better the series of tests necessary and to interpret their results correctly. In this context, co-operation between the laboratory doctor, the general practitioner and the gynaecologist is essential, especially to use resources more appropriately and to give patients correct information in order to be able to face an eventual pregnancy or understand the reasons behind previous failed attempts.

To summarise, the evidence linking thrombophilia to obstetric complications is weak and not always easy to interpret. Thromophilia appears to be common in women who have had otherwise unexplained recurrent abortions, with a variable incidence up to 65% in very selected populations. Carriers of thrombophilic alterations also have a higher risk of other complications attributable to vascular changes, although it is not possible to quantify the risk precisely until data from longitudinal studies become available. Despite the fact that the present data are concordant in showing an increased risk throughout the whole pregnancy, the risk is higher in the second and third trimesters than in the first. Thrombophilic disorders are not always convincingly associated with pre-eclampsia or IUGR. In fact, although thrombophilia is connected with up to 67% of gestational complications, the same thrombophilic disorders are also present in up to 20% of normal pregnancies, which suggests there must be other additional and necessary risk factors for the development of these complications.

Analytical procedures: are all the tests the same?

If identifying the patient is the first fundamental step in evaluating the pathway for diagnostic studies of thrombophilia in the context of pregnancy, the successive step is an evaluation of the type of investigations that should be used. Over the years, in particular after 1993⁴², the number of potential tests for studying thrombophilia has increased substantially. After thoroughly analysing changes in the concentrations or activities of the numerous coagulation factors and their inhibitors, the research is directed towards genetic polymorphisms, further emphasising the issues of whether the requests are appropriate and the significance of the interpretation of the results. In the context of laboratory medicine, it seems essential that there is very considerable rigour in the evaluation of the pre-analytical procedures (collection and management of the biological samples), analytical analyses (precise and reproducible methods) and post-analytical analyses (correct interpretation of the data) in order to guarantee a high standard of quality throughout the whole diagnostic process. From guidelines on the subject, standardisation of the pre-analytical procedures has clearly emerged as the determinant factor^198,199. For all coagulation tests, the pre-analytical stage represents the principal cause of error and imprecision as inadequate procedures contribute markedly to modifying the clinical context of the tests requested^200–205. The most critical elements are those related to the procedures for preparing the patient (fasting, biological variants), collection of the sample (equipment, test-tubes, relationship between blood and anticoagulants) and treatment of the sample (storage, centrifugation)¹⁹⁸. There are reliable data showing how failing to respect essential quality standards in the pre-analytical stage can have very negative repercussions on resource management and the patient’s outcome^206–213.

Another critical point is the correct interpretation of the results. A typical example is the method for evaluating APCR. There are numerous tests currently available and they have different analytical characteristics. The classical method, originally proposed by Dahlback et al.⁴³, evaluated APCR in toto. Other tests using FV-depleted plasma are specific for the FV Leiden mutation and have a sensitivity very close to 100%. However, these versions of the test are not able to identify the 10% of cases of APCR due to causes other than FV Leiden, such as elevated FVIII values, other FV polymorphisms, functional deficiencies of protein S and yet others. This example highlights the need for understanding between the clinician and laboratory analyst regarding the correct interpretation of the results. There are numerous other examples of how laboratory results can be interpreted differently in relation to the analytical and diagnostic efficiency (determination of lupus anticoagulant and antiphospholipid antibodies, protein C and protein S). It is necessary to be similarly rigorous regarding the period in which to carry out the analyses. Hormone replacement treatment, pregnancy and the use of oral contraceptives can have different effects on many parameters of blood coagulation, including protein S and APCR^134,214,215. Being aware of these changes is crucial both when requesting an investigation and when interpreting its results. The analytic problems with molecular biological diagnosis of FV Leiden, the prothrombin gene polymorphism, MTHFR, and other genetic polymorphisms are, however, less pronounced.

One still unresolved problem is the choice of the best battery of tests and diagnostic algorithms^123–126. The laboratory tests supported by the greatest clinical and epidemiological evidence seem to be those for protein C and protein S (free and functional forms), evaluation of APCR, lupus anticoagulant, and anticardiolipin antibodies as well as molecular biological investigations for FV Leiden, prothrombin and MTHFR. Other tests, particularly genetic ones, have not yet been sufficiently supported by scientific literature and should be reserved for selected cases with very suggestive clinical indications, in which the major causes of thrombosis and/or gestational complications have been excluded⁵¹. Finally, it should be remembered that the clinical management of patients with confirmed thrombophilia must be appropriate and never improvised. Close collaboration between different medical disciplines is necessary and patients should be cared for in highly specialised centres, in which the expertise and different medical skills can interact to provide the best diagnostic, prophylactic and therapeutic procedures.

Conclusions

The evaluation of thrombophilia screening in pregnant women is still a wide open field of research. On the basis of accredited scientific literature it currently seems unjustified to request a large battery of tests for all women of fertile age since this approach is not supported by a good cost-benefit ratio. That said, in the case of patients with previous thromboembolic episodes during pregnancy or gestational problems likely to be thrombotic in nature, the significant association with predisposing thrombophilic conditions could justify a search for the most frequent and serious alterations, such as the FV Leiden mutation, polymorphisms of the prothrombin gene, APCR, protein C and protein S deficiencies and hyperhomocysteinaemia. Given the complexity of the problem, the patient should be referred to a highly specialised centre able to provide the necessary counselling and the most suitable prophylaxis and treatment. In conclusion, further clinical and epidemiological research is necessary in order to clarify the still doubtful role of some inherited causes of thrombophilia in the context of pregnancy.