Prothrombin (PT, also known as clotting factor II) is a vitamin K-dependent proenzyme that induces the conversion of fibrinogen to fibrin, via a prothrombinase complex. PT was first reported by Loeliger in 1959 as a possible cofactor for LA. Since then, many other authors have suggested that antibodies binding to PT could contribute to LA phenomenon and, at the present time, they are largely considered as one of the most important causes of the elongation of clotting time due to LA positivity, together with anti-β₂GPI antibodies [6, 7].

Antibodies against prothrombin are usually detected by two different ELISAs (enzyme-linked immunosorbent assay) that employ human PT coated onto irradiated plates (aPT) or PT in complex with phosphatidylserine (aPS/PT) as antigen. These two assays seem to display different diagnostic and prognostic power, possibly attributable to their different ability to offer the antigen to antibody binding [6–8]. The real prevalence of aPT is still unknown, as it widely varies among studies, as a result of the variability of detection methods and the poor standardization among different laboratories. Also the clinical significance of aPT in both primary and secondary APS is still debated, as an apparent association with obstetric or thrombotic manifestations has not been definitely demonstrated [6]. On the other hand, aPS/PT antibodies strongly correlate with the presence of LA and are suggested to be highly specific for the diagnosis of APS [7, 8].

Most of the studies addressing the clinical significance of aPS/PT antibodies have demonstrated a significant association with thrombotic manifestations of APS, venous thrombosis above all, while the association with obstetric manifestations is still controversial. Actually, even if some authors reported that aPS/PT can be predictive of pregnancy morbidity in patients with systemic autoimmune diseases, most of the studies did not confirm this finding [9, 10].

Several studies have clearly demonstrated that multiple positive aPL tests are stronger predictor of thrombosis and pregnancy complications than single positivity. Accordingly, it has been suggested that the combination of routinely tested aPL (aCL, anti-β2GPI, and LA) with new (non-criteria) aPL assays would lead to a better risk stratification of patients. Among 23 possible combinations of six aPL assays (LA, aCL, anti-β2GPI, aPT, aPS/PT, and anti-phosphatidylethanolamine antibodies), the association of LA plus anti-β2GPI plus aPS/PT antibodies has recently been identified to display the best diagnostic accuracy for both vascular and obstetric APS [11].

A very recent and exhaustive review of the literature showed that both aPT and anti-PS/PT positivity increase the risk of thrombosis but that aPS/PT display the highest odds ratio (5.11 (95 % CI 4.2–6.3) vs 1.82 (95 % CI 1.44–2.75)). In line, the studies directly comparing aPT and aPS/PT and their odds ratio for thrombosis in 1196 patients demonstrate that aPS/PT antibodies are more strongly associated with both arterial and venous thrombosis than aPT antibodies [12].

Even if aPS/PT represent a very promising biomarker of APS, the lack of harmonization and standardization of the detection procedures and the low reproducibility of the results among laboratories are still unsolved problems. Promising data have been recently reported by Amengual et al. who have compared different assays for the detection of aPS/PT and found a good accuracy of both homemade and commercial ELISA kits and a high concordance of the results [13].

β2GPI is a large anionic plasma glycoprotein, consisting of 326 amino acids, organized in five domains [14]. This protein is highly immunogenic, and it has been demonstrated that autoimmune patients can produce antibodies against several epitopes of the molecule, located in different domains. Antibodies directed to different β2GPI domains seem to display higher or lower clinical significance [15]. Anti-domain I (DI) antibodies were firstly described in 2002, but their importance has clearly emerged more recently [16]. Actually, even if there is growing evidence that domain I represents the immunodominant epitope of β2GPI, the clinical significance of anti-DI antibodies is still debated. De Laat et al. have shown that anti-β2GPI antibodies with DI specificity are associated with LA and that anti-DI positivity correlates with vascular thrombosis, with an OR for venous thrombosis ranging from 3.5 to 6.7 in different studies [17, 18]. In one of these studies, the group found a correlation between anti-DI antibodies and obstetric APS manifestations, even though to a lesser extent than with thrombosis [18]. However, the results of this study have to be carefully evaluated, as it showed no correlation between LA and miscarriages, conflicting with several previous publications and the known clinical LA predictive value for miscarriages [19, 20].

Moreover in a recent study by our group, a high prevalence of anti-DI antibodies was detected in both thrombotic and obstetric primary APS, albeit anti-DI IgG were not found to be predictive of these complications [21].

Antibodies directed to the other domains of β2GPI seem to have lesser predictive value for APS. For example, anti-domain IV (DIV) and domain V (DV) antibodies have been found in patients with chronic infections, such as leprosy, in children with atopic dermatitis and in aPL-positive asymptomatic carriers [22].

Thus, definite conclusions on the diagnostic and prognostic value of anti-DI antibodies cannot be drawn at present, as the data regarding the association with thrombosis are not univocal among different studies [17, 18, 21]. Moreover, a small but relevant proportion of anti-β2GPI-positive APS patients do not display anti-DI antibodies, suggesting that the assay for the whole molecule cannot be substituted up to now [21].

It has been proposed that the ratio between anti-β2GPI-DI and anti-β2GPI-D IV/V IgG antibody reactivities could add important information to discriminate between relevant anti-β2GPI positivity associated with an autoimmune disease (such as APS) and antibodies occurring in association with other pathologies, with less diagnostic and pathogenic value. If confirmed in larger studies, this finding would suggest that tests for antibodies against the different domains could help in the risk stratification of anti-β2GPI antibody-positive patients [23].

There are different methodologies that can be employed to detect these antibodies. Besides the two-step assay, three ELISAs and a CIA using different DI molecules or peptides have been reported [23, 24].

Phosphatidylethanolamine (PE) is a neutral phospholipid that constitutes the inner leaflets of biological membranes. PE has anticoagulant properties and the finding that PE can interfere with clotting time prolongation raised the hypothesis that aPE might be responsible for the LA phenomenon, even if a significant association between aPE and LA has not been demonstrated [25].

In in vitro experiments, PE has been found to be an essential cofactor for the protein C anticoagulant pathway. Moreover PE is an inhibitor of the factor Xa-prothrombin system [26].

Antibodies targeting PE have been reported in up to 43 % of APS patients [25]. In a population of women with a history of recurrent early pregnancy loss, aPE prevalence has been reported to range between 23 and 31.7 % [27]. In another study, the rate of aPE positivity in a population of patients with otherwise unexplained thrombotic events was 18 % when detected by ELISA and 30.5 % when tested using thin-layer immunostaining [25, 28]. Moreover, in a multicenter study on 270 thrombotic patients, 63 % of 40 aPE-positive subjects had no additional aPL laboratory tests [25]. On the basis of these data, some authors have proposed aPE as serological markers of seronegative APS. However, given the limited number of studies, the small sample size, and poor ELISA standardization, the clinical role of aPE is still not clear and aPE testing is still not recommended.

The diagnostic and prognostic role of several autoantibodies against negatively charged PLs (other than CL) have been also studied. Among them, most data regard phosphatidylserine (PS), phosphatidylinositol (PI), and phosphatidic acid (PA). In the past, aCL has been shown to cross-react with antibodies targeting both PS and PI. Further studies have demonstrated that aPS, aPI, and aPA antibodies mainly recognize a complex consisting of β2GPI, coupled to these negatively charged aPL [28]. Therefore, most of the cross-reactivity is due to autoantibodies actually reacting with β₂GPI [5, 29].

There are no recent studies demonstrating that aPS, aPI, and aPA antibody testing significantly improves the diagnosis of APS [5]. Nevertheless, aPS antibody detection has been suggested to be useful in the context of pregnancy-related morbidity [29]. Recent available literature reports conflicting results. In one study aPS was not associated with recurrent pregnancy loss, while in another aPS IgG but not IgM were related to obstetric morbidity [5]. Moreover, in a cohort of women with a history of recurrent miscarriage, aPS was the only autoantibody detectable in 3.6 % of subjects [30]. The significant role of aPS in obstetric APS is also supported by in vivo studies on animal models. Two murine monoclonal antibodies targeting PS have been demonstrated to decrease human chorionic gonadotropin (hCG) secretion and to inhibit trophoblast invasion [30, 31]. Notably, one of these two aPS reacted with PS only, while the other was able to recognize both CL and PS, but no information about a possible cross-reactivity with β2GPI was available [30, 31]. More recently, active immunization with β2GPI-dependent aPS was able to induce fetal resorption in a murine model of APS [5]. However, data on humans are inconsistent and aPS assays are still not included in classification criteria.

Recently, a novel aPL assay (APhL IgG/IgM ELISA), using a mixture of negatively charged phospholipids as antigen (including PS, phosphatidic acid, and β2GPI), has been tested in APS patients [32]. Positivity for this commercial ELISA kit has been reported to be more predictive for APS than aCL. Particularly, APhL test showed higher positive and negative predictive values for APS in comparison to two commercially available aCL assays. Moreover, in the same study, antibodies against APhL have been associated with arterial events in a cohort of SLE patients. The authors suggest that the routine use of this assay could implement specificity, without losing sensitivity for APS [32].

Annexins are a family of proteins that bind Ca⁺⁺ and phospholipids and display several different functions, including inhibition of coagulation processes in the vasculature and on trophoblasts [33]. Annexin V (AnnA5) is present on the intervillous surface of the placenta, forming a shield that prevents the activation of the coagulation cascade [33, 34]. Several studies have reported a significant reduction of annexin V binding on the placental tissue from patients with obstetric APS in comparison with normal controls. In addition, aPL have been shown to displace annexin V from both trophoblast and endothelial cell monolayers in in vitro studies [33, 34].

Recently, the determination of resistance to the anticoagulant effects of AnnA5 (AnnA5 resistance) has been proposed as a marker of APS. AnnA5 resistance has been found in a significantly higher proportion of APS patients in comparison to controls. Moreover, a significant reduction of AnnA5 anticoagulant activity was detected in a cohort of SLE children in comparison to pediatric controls [35]. Notably, the same cohort of patients displayed a significant increased prevalence of anti-DI antibodies (p = 0.014) compared to controls, and resistance to AnnA5 anticoagulant activity was found to inversely correlate with titers of anti-DI IgG antibodies [35]. Even if these promising data suggest that AnnV resistance could play a role in the identification of specific subsets of pathogenic aPL antibodies, further studies are needed to confirm this preliminary finding.

Annexin 2 (AnnA2) has been proposed as a target of aPL. Several studies have suggested that AnnA2 could represent a receptor mediating β2GPI binding to endothelial cells (ECs) [36]. Autoantibodies against AnnA2 have been described in patients with APS and severe thrombosis and/or pregnancy morbidity but also in some other autoimmune conditions (such as SLE and RA). Their clinical significance is unclear at the moment [36].

The clinical significance of aCL and anti-β2GPI antibodies of the IgA isotype in PAPS is still a controversial issue. IgA aCL and/or IgA anti-β2GPI antibodies have been reported in seronegative patients with a history of thrombosis and pregnancy morbidity. Particularly, IgA anti-β2GPI antibodies can potentially identify APS in patients who possess the clinical features of the disease but do not meet current laboratory criteria. In mouse models these antibodies were able to induce significantly larger thrombi and higher tissue factor levels compared to controls, demonstrating their pathogenic role. In a recent study of Mattia et al. on 84 PAPS patients, IgA aCL and IgA anti-β2GPI antibodies were found, respectively, in 19 and 50 % of patients. The mean titers of both IgA aCL and IgA anti-β2GPI antibodies were higher in the thrombotic patients, but only IgA anti-β2GPI were significantly associated with thrombosis. Isolated IgA anti-β2GPI antibody positivity was significantly prevalent in seven of the seronegative patients [37]. There are several reasons to explain why a number of studies failed to prove the usefulness of adding IgA aCL and IgA anti-β2GPI testing. In fact, these autoantibodies have a low prevalence and are mostly found in association with other aPL. Moreover, few accurate diagnostic tests are available for their detection [38]. Recently, Ruiz-García et al. found that mean levels of IgG, IgM, and IgA, both aCL and anti-β2GPI, antibodies were significantly higher in patients with clinical features of APS than in controls on a total of 156 patients fulfilling clinical criteria for APS. IgA anti-β2GPI was the most prevalent antibody in these patients [39].

In addition to thrombosis and pregnancy morbidity, a number of clinical manifestations have been described in aPL-positive patients. Non-thrombotic neurological features, thrombocytopenia, heart valve disease, microangiopathic nephropathy, livedo reticularis, and skin ulcers are some of the possible features that are not yet considered as classification criteria because of their low specificity (Table 6.2).