The phenomenon known as acquired activated protein C resistance

Resistance to the effect of activated protein C (APC), added to plasma from patients with a history of deep-vein thrombosis, was reported by Amer et al (1990), and the significance of inherited resistance was recognized by Dahlback et al (1993). In the majority of subjects, inherited resistance results from the Leiden mutation (factor V Leiden, FVL) in the gene coding for coagulation factor V (Bertina et al, 1994), although a number of other genetic variations have been described in the factor V gene associated with an APC resistance phenotype. Inherited resistance has been identified as a significant risk factor for venous thrombosis (Griffin et al, 1993; Koster et al, 1993), late fetal loss (Martinelli et al, 2000) and pre-eclampsia (Dizon-Townson et al, 1996). In recent years, the potential significance of resistance occurring in the absence of known mutations of the factor V gene has been recognized. A number of authors have coined the term acquired APC resistance for this phenomenon. The investigation of APC resistance by Dahlback et al (1993) was based on the effect of APC on an activated partial thromboplastin time (aPTT). Since then, it has been recognized that the originally described aPTT-based test is not specific for the Leiden mutation. Indeed, in the intervening years, many strategies have been devised to render coagulation-based APC resistance assessments more specific for FVL (de Ronde & Bertina, 1994; Jorquera et al, 1994; Varadi et al, 1995; Nicolaes et al, 1996). Despite this drive to improve specificity, studies that used the original aPTT-based test or techniques that remain sensitive to a variety of causes of APC resistance (Gable et al, 1997; Nicolaes et al, 1997) have shown that the degree of resistance (in FVL-negative subjects) relates to: • thrombin generation (Clark et al, 1999; Lowe et al, 1999); • venous thrombosis risk (de Visser et al, 1999); • cerebrovascular disease risk (van der Bom et al, 1996; Kiechl et al, 1999); • pre-eclampsia risk (Clark et al, 2001a); • the presence of antiphospholipid activity (Hampton et al, 1994); • combined oral contraceptive use (Osterud et al, 1994; Henkens et al, 1995; Olivieri et al, 1995), in particular third-generation oral contraceptive use (Rosing et al, 1997). A substantial number of questions are posed by these observations. These are typified by the debate resulting from the observation that use of the third-generation combined oral contraceptive pill (COCP) is associated with more APC resistance than second-generation COCPs (Rosing et al, 1997). The authors used an APC resistance test based upon the thrombin generation resulting from tissue factor-activated plasma. However, a comparable differential of resistance has not been seen with aPTT-APC resistance tests (Henkens et al, 1995; Olivieri et al, 1995; Schramm & Heinemann, 1997), a finding that has subsequently been confirmed by Rosing's own group (Curvers et al, 1999). Although this variance has highlighted the difficulty of diagnosing acquired APC resistance, it is likely to afford an opportunity to understand the mechanism of the additional risk of venous thrombosis (venous thromboembolim, VTE) associated with third-generation COCPs (Vandenbroucke & Rosendaal, 1997). From this debate, it is clear that the detection of acquired APC resistance requires consideration of the method of testing and the population under study. The APC resistance phenotype does, however, correlate with markers of in vivo thrombin generation (Freyburger et al, 1997; Clark et al, 1999; Lowe et al, 1999) and may indicate an overall prothrombotic phenotype. What information, then, does APC resistance testing give us and, given the myriad of influences upon it, can APC resistance testing (in non-FVL subjects) be tamed to provide a useful tool in the assessment of thrombotic risk? In the assessment of APC resistance, distinction has to be drawn between factor V mutation-specific and non-specific assessments. In the main, factor V-specific assessments involve predilution of subject plasma. This may be achieved with factor V-depleted plasma (Jorquera et al, 1994), which allows the cofactors of the protein C/protein S system to be supplied by the added normal plasma. The dilution of factor V renders the test highly specific to the presence of factor V mutations. Similar strategies include predilution of the patient plasma before assessment of the effect of APC on factor VIIIa/factor Xa generation (Varadi et al, 1995), the effect of APC on factor Va generation (assessed in a prothrombinase system) (Nicolaes et al, 1996) or the effect of APC on a modified factor V assay (Le et al, 1995). These strategies exclude inhibitors and anticoagulants, but also render the test insensitive to other prothrombotic influences. In addition to the original aPTT test, other methods of APC resistance assessment have been described that remain sensitive to a variety of influences (Kraus et al, 1995; Dati et al, 1997; Nicolaes et al, 1997). In particular, the assessment (described above) by Nicolaes et al (1997) examines the effect of APC on the generation of thrombin (measured by the formation of thrombin–α2-macroglobulin complexes with time) from tissue factor-activated plasma. In the aPTT-APC resistance test described by Dahlback et al (1993), a ratio of the aPTT obtained in the presence of added APC/the aPTT obtained without added APC is derived. With this APC sensitivity ratio (APC:SR), a low ratio indicates high resistance. A number of coagulation factors influence both the aPTT (Dahlback et al, 1993) and tissue factor (Nicolaes et al, 1997) assessments. aPTT-APC resistance correlates positively with factor VIIIc levels during pregnancy (Clark et al, 1998), in association with markers of an inflammatory state (Laffan & Manning, 1996; Mathonet et al, 1996), and in the population as a whole (Lowe et al, 1999). This is consistent with the association of both factor VIIIc (Koster et al, 1995; Kraaijenhagen et al, 2000) and blood group (Talbot et al, 1972; Meade et al, 1994) with thrombosis. aPTT-based resistance also correlates positively with factor IXc (Lowe et al, 1999), Xc and XIc levels (Kluft et al, 1999). This may, in part, be confounded by the relationship of these variables to each other and the aPTT (Lowe et al, 1997; van Hylckama Vlieg et al, 2000), although factor IXc has been recognized as an independent risk factor for venous thrombosis (van Hylckama Vlieg et al, 2000). In pregnancy, aPTT-APC resistance also correlates positively with factor Vc (Clark et al, 1998), but this influence is only observed at extreme factor Vc levels in non-pregnant subjects (Colucci et al, 1994; Freyburger et al, 1996). An inverse relationship to protein S is seen in pregnancy (Clark et al, 1998), and higher resistance is observed in protein S-deficient subjects (de Ronde & Bertina, 1994). However, no influence on resistance is apparent at ‘normal’ protein S levels (de Ronde & Bertina, 1994; Freyburger et al, 1997; Lowe et al, 1999). No relationship with ‘normal’ levels of protein C has been reported (Freyburger et al, 1997; Lowe et al, 1999), but increasing resistance is associated with increasing antithrombin levels (Freyburger et al, 1997; Lowe et al, 1999). This may indicate a compensatory response to a prothrombotic state. aPTT-APC resistance is also associated with lupus inhibitors (Bokarewa et al, 1994). The observation of higher resistance associated with anti-B2 glycoprotein-1 antibodies and anticardiolipin antibodies (Martinuzzo et al, 1996) suggests that this effect is not solely in vitro interference with the aPTT. Indeed, anticardiolipin antibodies can inhibit the action of APC on factor Va (Oosting et al, 1993), and a dose–response relationship between antiphospholipid-related resistance and thrombosis has been suggested (Aznar et al, 1997). Like the aPTT test, the tissue factor-APC resistance test is affected by lupus inhibitors and is unreliable in the presence of anticoagulants (Nicolaes et al, 1997). Protein S deficiency is also associated with higher tissue factor-APC resistance (Nicolaes et al, 1997). In pregnancy, a gestation-independent positive relationship of resistance with cholesterol, antithrombin and protein C activity and an inverse relationship with total protein S has also been reported (Clark et al, 2001b), although there is little other published information on the influence of coagulation factors on this test. Heightened APC resistance, as measured by an aPTT-based method, was reported in association with COCPs containing ≤ 37·5 µg of ethinyl oestradiol (Henkens et al, 1995). In this study, no distinction between second-generation (levonorgestrel or norgestimate) and third-generation progestogens (desogestrel or gestodene) was made. The bulk of subsequent studies has shown that third-generation progestogens are associated with a 1·4- to fourfold increased risk of VTE compared with second-generation COCPs (reviewed by Vandenbroucke et al, 2001). As noted above, use of the third-generation COCP is associated with more APC resistance than second-generation COCPs (Rosing et al, 1997), when estimated by the thrombin generation resulting from tissue factor-activated plasma. Although not seen with aPTT-based APCR assessments, this phenomenon has been offered as an explanation for the additional risk of VTE associated with third-generation COCPs (Vandenbroucke & Rosendaal, 1997). Consistent with observations of COCPs, recent epidemiological studies have suggested that the use of hormone replacement therapy (HRT) is associated with a threefold relative risk of deep-vein thrombosis (DVT) (Chae et al, 1997; Meade, 1997). A comparable increase in aPTT-APC resistance to that observed with oral contraceptive use has been observed in HRT users (Lowe et al, 1999), a relationship that may be independent of factor VIIIc and IXc (Nebulsi et al, 1993; Woodward et al, 1997). A number of demographic variables also influence the aPTT-APC resistance test (Table I). These reflect the positive influence of oestrogen, factor VIII and acute-phase reactants on APC resistance (Fig 1), and the majority are also in keeping with the concept of APC resistance as a phenotypic marker of venous and arterial disease risk. The relationship between smoking and a reduction in APC resistance appears, however, to be incongruous. There is, indeed, conflicting evidence that smoking is a risk factor for VTE (Marks & Emerson, 1974; Emerson & Marks, 1977; Prescott et al, 1978; Goldhaber et al, 1997). However, the relationship of smoking to APC resistance is consistent with the mitigating effect of smoking on the risk of pre-eclampsia (van der Velde & Treffers, 1985; Klonoff-Cohen et al, 1993). Serum lipid levels are an accepted risk factor for arterial disease but are not generally considered to be a risk factor for venous thrombosis. Small preliminary studies have, however, suggested a link between serum lipids, thrombin generation (MacCallum et al, 2000) and VTE (Kawasaki et al, 1995, 1997). Potential mechanisms of acquired APC resistance. The potential link between demographic factors and APC resistance (APCR) via alterations in coagulation factor levels is shown. FVc, factor Vc; PS, protein S; FVIIIc, factor VIIIc; HRT, combined hormone replacement therapy. A potential linkage from these factors via APCR and thrombin generation to pre-eclampsia (PET), venous thrombosis (VTE) and stroke (CVA) is outlined. Higher factor VIIc levels (FVIIc) are associated with higher triglyceride (TRIG) levels. The mechanism of triglyceride-related APCR is unknown. Only the additional coagulation effect of third-generation combined oral contraceptives (third-generation COCPs) over second-generation COCPs is shown. Third-generation COCPs are associated with higher tissue factor-APC resistance (not aPTT-APCR) and higher FVIIc levels than second-generation COCPs. The mechanism of higher tissue factor-APC resistance is unknown, but there may be a direct effect of FVIIc. #A link between factor V Leiden (FVL), factor V Cambridge (FVC), or the HR2 haplotype (HR2) and cerebrovascular disorders is not proven. Letters indicate the source publication: (Tosetto et al, 1997)a; (Clark et al, 2001a)b; (Clark et al, 1998)c; (Rosing et al, 1997)d; (Lowe et al, 1999)e; (de Visser et al, 1999)f; (Laffan & Manning, 1996)g; (Mathonet et al, 1996)h; (Clark et al, 1999)I; (van der Bom et al, 1996)j; (Kiechl et al, 1999)k; (Bertina et al, 1994)l; (Williamson et al, 1998)m; (Bernardi et al, 1997)n; (van Baal et al, 2000)°; (reviewed by Winkler, 1998)p. Given the number of potential confounders (Table I), it is perhaps not surprising that only a few studies have attempted to determine whether acquired APC resistance is an independent risk factor for thrombotic disease. These studies have avoided the use of artificial cut-offs to characterize subjects as having APC resistance and have considered resistance as a relative phenomenon. In 1996, a report from the Rotterdam Study described a 2·6-fold (95% CI 1·2–5·6) increased risk of cerebrovascular disease when subjects with the highest quintile of resistance were compared with those with the lowest (van der Bom et al, 1996). No significant difference in the risk was observed when the data were corrected for age, sex, cholesterol level and FVL status. In a population study by Kiechl et al (1999), a dose–response relationship between quintiles of resistance and the occurrence of femoral artery and carotid artery stenosis was observed. When adjusted for sex, menopause status, smoking and factors related to factor VIII levels (age, the aPTT and acute phase-reactants), a strengthening of the relationship was observed. Furthermore, no significant difference in the three- to fourfold risk of stenoses was observed when the data were adjusted for the presence of the Leiden mutation. As noted above, smoking is associated with lower APC resistance and a reduced risk of pre-eclampsia (PET). In a study of FVL-negative subjects, the degree of resistance in the first trimester of pregnancy related to the risk of development of pre-eclampsia in the index pregnancy (Clark et al, 2001a). A sevenfold risk of pre-eclampsia was observed for those in the highest quartile when compared with those in the lowest. In this study, regression for those factors that confounded the relationship (smoking and booking blood pressure) revealed a persisting relationship between APC resistance and the risk of subsequent PET. The confounding of aPTT-APC resistance by factor VIIIc levels was addressed directly by de Visser et al (1999). In FVL-negative, lupus inhibitor-negative subjects, the highest quartile of resistance was associated with a fourfold risk of venous thrombosis when compared with those subjects in the lowest quartile. Forty-five per cent of those subjects in the highest quartile had a factor VIIIc level > 150 iu/dl. Although adjustment for FVIIIc levels, age and sex did not abolish the relationship with thrombosis, the relative risk associated with the highest quartile was reduced to 2·5 (95% CI 1·5–4·2). Thus, APC resistance appears to be an independent marker of a prothrombotic phenotype. The relationship with such a wide variety of prothrombotic influences (Table I, Fig 1), markers of in vivo thrombin generation (Clark et al, 1999; Lowe et al, 1999) and thrombotic disease (van der Bom et al, 1996; de Visser et al, 2000; Clark et al, 2001a) suggests that APC resistance is a common mechanism whereby a number of risk factors for arterial and venous thrombosis promote thrombogenesis. Could this disparate phenomenon usefully be used as a screening tool to determine risk? A number of questions would have to be answered before this could be considered: • What type of test should be used? Clearly, each resistance assessment gives different demographic and coagulation information. At present, there is greater experience in the use and interpretation of the aPTT-based test. It would seem likely that this would become the most widely applicable acquired APC resistance screening tool. • How should testing be standardized? Standardization has already been recommended for aPTT-APC resistance screening for FVL (de Ronde & Bertina, 1994) and, given the variety of influences upon it, an aPTT test for acquired resistance would require population-specific reference ranges. • When is the appropriate time to test? Some resistance relates to acute-phase reactants, and the link between this and future thrombotic risk is unknown. Furthermore, testing for acquired resistance can only be achieved on non-anticoagulated patients. More importantly, in addition to these technical considerations, a number of fundamental principles of screening require to be addressed: • Will screening alter management? It is not yet clear whether this question has been answered for heritable resistance (Ridker et al, 1995; Rintelen et al, 1996; Simioni et al, 1997; Eichinger et al, 1999; Kearon et al, 1999). Elevated levels of in vivo markers of thrombin generation are seen in heritable APC resistance (Martinelli et al, 1996; Simioni et al, 1996), and these markers may be reduced by anticoagulation (Takahashi et al, 1993; Koefoed et al, 1997). Owing to the attendant risk of bleeding, the use of anticoagulants requires a subtle balance of risks and benefits. Unlike FVL, acquired APC resistance does relate to a number of thrombotic risk factors, such as cholesterol, blood pressure and body mass index (BMI) (Table I). It is conceivable that intervention to reduce these thrombotic risk factors could offer additional modalities of therapy to reduce APC resistance and thrombotic risk. • How is an abnormal resistance defined? A number of studies have shown that there is a dose–response relationship between the degree of resistance and the development of disease. That the risks of vascular disease (in those with the lowest quartile/quintiles of resistance) are comparable with those observed in heritable resistance is remarkable, given that the degree of resistance is much less. Intervention to reduce disease would clearly require a new definition of abnormal or unacceptable resistance. The move from progressive risk to action limits is, thankfully, not new in the field of vascular disease. The discovery of heritable APC resistance has transformed our understanding of VTE. The subsequent discovery of corresponding defects in the factor V gene has resulted in a concentration on refinement of coagulation tests to improve their specificity and sensitivity for these mutations. In recent years, the significance of non-Leiden resistance has been realized. Much work is still required to determine whether this prothrombotic phenotype usefully can be added to routine thrombophilia screening.