Direct oral anticoagulants (DOACs) include the direct thrombin inhibitor dabigatran etexilate and the direct anti-factor Xa inhibitors rivaroxaban, apixaban, and edoxaban. These agents, unlike warfarin, are prescribed in fixed doses with more predictable anticoagulant effects, do not interact with diet or alcohol, and have fewer reported drug interactions that affect anticoagulant intensity. Furthermore, monitoring of anticoagulant intensity of a DOAC is not routinely required because anticoagulant effects are more predictable.

The 14th International Congress on aPL Task Force on Treatment Trends concluded that “warfarin or other VKA remains the mainstay of anticoagulation in thrombotic APS. Direct oral antiocagulants can be considered in APS patients with a first or recurrent venous thromboembolism (VTE) occurring off or on subtherapeutic anticoagulation, only when there is known VKA allergy/intolerance or poor anticoagulant control. There were no data to recommend DOACs in APS patients with recurrent VTE occurring on therapeutic anticoagulation or with APS-related arterial thrombosis” [3].

A systematic review (through April 2016) presented during the 15th International Congress on aPL identified seven case reports and four case series . They included 99 DOAC-treated APS patients, of whom 38 had primary APS, 23 APS associated with lupus, and 38 unspecified. Approximately 20% of patients had vascular events during a mean follow-up of 12 months [4]. More recently, in a prospective case series, six (four triple aPL-positive) of 56 (11%) APS patients developed recurrent thrombosis on DOAC (mean follow-up 22 months) [5].

The anecdotal clinical reports (in which rivaroxaban was used in the majority of patients), with recognition of their inherent limitations of publication bias and low evidence level study designs, suggest that recurrent thrombotic events with DOACs in APS patients mainly occur when DOACs are used for secondary prevention of aPL-related arterial thrombosis or microthrombosis, situations where DOACs are not approved. They highlight the need for randomized controlled trials to guide the use of DOACs in thrombotic APS. One recently completed and two ongoing controlled clinical trials and a feasibility study on DOACS were presented during the Congress.

Benefit of statins in primary and secondary prevention of coronary heart disease is proven, which is due to the lipid-lowering effect of these drugs and to their immunomodulatory, anti-inflammatory, and antithrombotic properties [9]. Statins have multiple effects on monocyte, lymphocyte, and endothelial cell activities that may contribute to thrombosis prevention in APS. Antiphospholipid antibodies induce expression of tissue factor (TF) and cell adhesion molecules; fluvastatin, simvastatin, and rosuvastatin reduce this expression [10]. In a mouse model of obstetrical APS, simvastatin and pravastatin reduced fetal death by inhibiting TF and protease-activated receptor 2 (PAR2) expression on neutrophils [11]. In a thrombosis mouse model, fluvastatin reduced thrombus size [12].

Based on in vitro and two human studies using surrogate markers [13, 14], the 14th International Congress on aPL Task Force on Treatment Trends concluded that “although statins ameliorate the proinflammatory profile and down-regulate the prothrombotic stage found in APS patients, based on the available data, statins cannot be recommended in APS patients in the absence of hyperlipidemia. However, a subgroup of aPL-positive patients with recurrent thrombosis despite adequate anticoagulation might derive benefits from statins [3].”

Since then there have been no systematic studies with statins in thrombotic APS. A study presented at the 15th International Congress on aPL [15] reported 21 pregnant APS patients who developed preeclampsia (PE) and/or IUGR on low-dose aspirin (LDA) and low molecular weight heparin (LMWH). Of those studied, 11 received pravastatin (20 mg daily) initiated at the onset of PE and/or IUGR (in addition to LDA + LMWH) and 10 did not. The study was not randomized, and not all patients met APS criteria at the time of enrollment. All pravastatin-treated patients had live births near full term. In the 11 patients who did not receive pravastatin, 11 deliveries were preterm, and five neonates died, and three of the six survivors had abnormal development. Patients treated with pravastatin had increased placental blood flow and improvements in PE features as early as 10 days after treatment was begun.

The rapid improvement in uterine artery hemodynamic parameters in pravastatin-treated patients (uteroplacental perfusion was assessed by Doppler) suggests that the drug targets placental vasculopathy, possibly by stimulating release of vasoactive substances, such as nitric oxide, from the endothelium [16]. Statins also downregulate TF, the major initiator of the coagulation cascade in vivo and a crucial molecule linking inflammation and thrombosis in APS [17]. These protective effects may explain the amelioration of placental and maternal PE [18].

Animal models demonstrate that B cells play an important role in the pathogenesis of APS. Blocking B-cell-activating factor (BAFF) prevents disease onset and prolongs survival in APS murine models [23]. Rituximab and belimumab are the only B-cell-inhibiting biologic therapies whose effect has been studied in APS patients.

Based on the case reports [23–25], CAPS registry data [26], and one open-label uncontrolled pilot study [27], the 14th International Congress on aPL Task Force on Treatment Trends concluded that “B-cell inhibition may have a role in difficult-to-treat APS patients, possibly in those with hematologic and microthrombotic/microangiopathic manifestations [3].”

Since then, although there have been case reports/series of rituximab use in APS and a recent report of belimumab use in two primary APS patients [28], there is not yet a prospective study of B-cell inhibition in APS.

Complement activation contributes to thrombosis in APS animal models. Antiphospholipid antibodies activate complement factor 5 (C5), generating fragment 5a (C5a), which induces adhesion molecules and TF in inflammatory cells and platelets and triggers release of prothrombotic and pro-inflammatory mediators [29, 30]. In animal models of APS, C5a interacts with its receptor (C5aR), amplifying endothelial cell activation and vascular inflammation, promoting trophoblast injury and angiogenic factor imbalance, and producing microthrombotic lesions [29, 31, 32]. Inhibition by anti-C5 antibody, C5aR antagonist peptides, and anti-C5aR antibody protects mice from pregnancy losses and from thrombosis [33]. Mice treated with aPL but deficient in complement regulatory components, i.e., mice in which the complement cascade is not normally inhibited, have poor pregnancy outcomes, including thrombotic microangiopathy. Mice treated with aPL, but deficient in complement components, have good pregnancy outcomes. These contrasting experiments indicate that complement is an important mediator in pathogenesis of both manifestations of APS [34]. Heparin, used in APS patients to treat thrombosis and prevent miscarriages, has a complement inhibitory effect [35].

Given the reports that complement inhibition improves outcomes in mouse models and CAPS patients [36], the 14th International Congress on aPL Task Force on Treatment Trends concluded that “complement inhibition may have a role as an adjuvant or main therapy for APS patients refractory to anticoagulation; however more clinical data are needed before this medication can be recommended” [3].

Since then a phase IIa study of treatment of non-criteria manifestations of APS (nephropathy, thrombocytopenia, and skin ulcers), designed to evaluate safety of an intravenous C5a inhibitor (clinicaltrials.​gov #: NCT02128269), has ended, but the results are not yet available. Another phase II study using eculizumab (a C5 inhibitor) to prevent thrombosis after renal transplantation in patients with prior CAPS (clinicaltrials.​gov#: NCT1029587) is currently recruiting.

Eculizumab is a fully recombinant humanized hybrid IgG2/IgG4 monoclonal antibody that binds to C5 and inhibits C5 cleavage by C5 convertase, thereby inhibiting the inflammatory, thrombotic, and lytic functions of complement. Eculizumab is FDA-approved for paroxysmal nocturnal hemoglobinuria (PNH) , an illness in which red blood cells and leukocytes are prone to complement-mediated lysis due to deficiency in CD55 or CD59, and for atypical hemolytic-uremic syndrome (HUS) [37]. A case report described complement inhibition in a refractory patient with CAPS who underwent kidney transplantation [38], and several other cases have been published [36]. Kidney allograft survival improves if complement blockade is initiated at the time of transplantation [39], suggesting a use in patients with APS nephropathy undergoing renal transplantation.

Complement plays a role in ischemia/reperfusion injury, antibody- and cell-mediated transplant rejection, C3 glomerulopathy, and atypical HUS [40]. Eculizumab was used to block thrombotic microangiopathy (TMA) in a patient with atypical HUS [41]. Three patients with APS (two with CAPS) treated with anticoagulation and eculizumab prior to and after renal transplantation had functioning renal allografts, with no systemic thrombotic events or early graft losses [42], confirmed in a report demonstrating improvement in TMA after eculizumab in APS nephropathy that recurred after kidney transplantation. In this case pretreatment intense C5b-9 and C4d deposit in kidney biopsies were absent after treatment with eculizumab, but chronic vascular renal changes were not prevented [43]. A recent case report described the potential use of eculizumab to prevent re-thrombosis in an arterial bypass graft in APS. Immunofluorescence showing β₂-glycoprotein-I (β₂GPI) on the endothelium of the artery wall suggested a pathogenic role for aPL [44].

Rivaroxaban, a direct factor Xa inhibitor, decreases markers of complement activation (C3a, C5a, and soluble [s] C5b-9) [45]. Antiphospholipid syndrome patients had higher baseline C3a, C5a, and SC5b-9 and, after 42 days of rivaroxaban treatment, a statistically significant decrease was observed (compared to controls treated with warfarin), possibly because rivaroxaban inhibits FXa-induced complement activation of C5 [46]. Further studies are necessary to confirm these findings.

Recently a novel recombinant antibody recognizing the first domain of β₂GPI was shown to induce thrombosis and fetal loss in animal models [47]; a non-complement fixing CH2-deleted variant of that antibody reduced the detrimental effects, suggesting another complement-mediated mechanism and another possible treatment for refractory APS [47].

All proposed peptide therapies are based on the premise that the key pathogenic interaction in patients with APS occurs when aPL engages one of the five domains (DI–DV) of β₂GPI [48]. Antibodies to several domains occur in APS patients, but most pathogenic aPL bind the N-terminal DI [49]. Each antigen-binding arm of the antibody can bind a separate β₂GPI molecule, creating a dimeric structure that binds anionic PL on cell membranes, thus interacting with cell surface receptors like Toll-like receptors (TLR) and apolipoprotein E receptor 2 (ApoER2) to alter cell function [50]. In the absence of aPL, β₂GPI is monomeric, is present constitutively in human serum, and does not exert these effects. The peptide agents proposed as new therapies for APS all act by blocking binding at different points. It is unknown which approach is best.

The 14th International Congress on aPL Task Force on Treatment Trends task force recommended that “at present, peptide therapy is not ready for trials in patients; however peptide therapy is potentially an important future targeted treatment for aPL-positive patients. Chemical modification to improve half-life and minimize immunogenicity will be required. Different peptides may be needed for different aPL manifestations” [3].

Since then, no peptide therapies have yet entered human trials. Considerable information from in vitro and animal studies suggests that peptides may become therapeutic agents in the future.

It is unknown which domain of β₂GPI may be the best therapeutic target. The key epitope for aPL on DI lies between the arginine residues R39 and R43 [49, 51]. In the in vivo femoral vein thrombosis model developed by Silvia Pierangeli, some DI peptide analogues inhibited thrombosis, TF expression in peritoneal macrophages, and aortic VCAM-1 expression, whereas others did not [52].

Linear DI peptides containing the critical R39–R43 epitope do not bind APS-IgG as well as does whole DI [53, 54]. McDonnell et al. have published a method for producing recombinant DI in bacteria [55] and have improved its half-life by adding polyethylene glycol group PEGylation [56, 57].

An alternative idea is to block DV binding to phospholipids. The octapeptide CKNKEKKC inhibits binding aPL to cardiolipin [58]. A 15-mer peptide from DV (called GDKV) and a cytomegalovirus peptide (TIFI) homologous to GDKV have been used as inhibitors. In mouse models that use injected aPL, TIFI inhibits thrombosis in the femoral vein [59] and reduces fetal loss [60]. TIFI also reduces binding of β₂GPI to human umbilical vein endothelial cells (HUVEC) [59] and to human trophoblast cells [60].

Blank et al. joined two synthetic peptides derived from DV (one was GDKV) with a flexible linker molecule [61]. When HUVEC were preincubated for a short while with the construct, binding of β₂GPI/anti-β₂GPI antibody (aβ₂GPI) complexes to HUVEC was reduced by 89%, but the reduction in binding was lost with prolonged preincubation because the construct was taken up in the cells, thus tempering enthusiasm for use of this agent.

Kolyada et al. developed a dimer, the A1 ligand-binding module of the ApoER2 receptor, that binds DV of β₂GPI [62–64]. Their construct blocks binding of both β₂GPI/ aβ₂GPI and DV alone to cardiolipin far more strongly than does monomeric A1 but does not do so in the absence of aβ₂GPI; it reduces thrombosis induced by laser trauma in two mouse models [63]. Recently the same group developed a mutant dimer that inhibits binding and clotting more strongly than does the wild type [64].

Recombinant DI, TIFI, and A1-A1 are all credible peptide therapies for APS. Proponents of TIFI and A1-A1 could argue that the heterogeneity of the aβ₂-glycoprotein-I in patients with APS mitigates against an approach that targets DI alone. Conversely, others might argue that interfering with the interaction of DV with PL may have adverse effects on physiological processes that occur in the absence of aPL. Perhaps there is a need for different peptide therapies. There is little information about potential toxic effects of any of these agents. All putative peptide therapies will have to be modified to enhance half-life [63].

A growing body of evidence highlights vitamin D’s immunomodulatory properties. Vitamin D insufficiency (<30 ng/ml) occurs in up to 70% of patients with APS and/or SLE, and vitamin D deficiency (<10 ng/ml) occurs in 11–50%. Low vitamin D levels in APS patients correlate with venous and arterial thrombosis and with non-criteria manifestations. Values in thrombotic APS patients are lower than in obstetric APS patients [65–67]. However, low-dose short-term vitamin D supplementation in a small group of primary APS patients was ineffective in raising levels above 30 ng/ml [67]. Human umbilical vein endothelial cell cultures demonstrate that vitamin D inhibits expression of prothrombotic TF in response to aβ₂GPI stimulation [65], a possible mechanism for therapeutic effect.

The 14th International Congress on aPL Task Force on Treatment Trends task force recommended that “vitamin D deficiency and insufficiency should be corrected in all aPL-positive patients based on the general population guidelines. The prognostic role of vitamin D deficiency and therapeutic value of supplementation (including the dosage and definition of treatment goals) in aPL-positive patients should be clarified with prospective studies that include appropriate control groups and standardized definitions of vitamin D deficiency” [3].

There have been no ongoing studies evaluating the effect of vitamin D in aPL-positive patients. However, a randomized clinical trial of vitamin D prophylaxis in the prevention of hypertensive disorders of pregnancy (clinicaltrials.gov #: NCT02920593) recently began. Investigators will also evaluate placental pathology and measure placental levels of several proinflammatory markers.

In the early stages of pregnancy, trophoblasts respond to and produce vitamin D, promoting an anti-inflammatory environment and inducing decidualization [68–70]. Vitamin D deficiency is associated with an increased risk for preeclampsia and IUGR [71]. A retrospective cross-sectional study of women with recurrent pregnancy loss (RPL) stated that vitamin D deficiency and insufficiency are associated with aPL, elevated peripheral blood natural killer (NK) cells, and elevated NK cell cytotoxicity [72]. In vitro studies confirmed the ameliorative effect of vitamin D on NK cell cytotoxicity and Th2-type response, both associated with successful pregnancy outcomes [72, 73].

Studies utilizing a human first trimester trophoblast cell line and primary trophoblast cultures show that vitamin D treatment alone or combined with LMWH limits aPL-mediated inflammatory response [74]. Vitamin D also reduces the elevated anti-angiogenic factor sFlt-1 seen in aPL- and/or LMWH-treated trophoblasts and placental villi explants [74–76]. The importance of this finding lies in the strong association of sFlt-1 with preeclampsia. Elevated levels of sFlt-1 are also seen in pregnant women treated with LMWH [77–78], possibly explaining the contradictory results of studies on the effectiveness of LMWH and aspirin in aPL-associated pregnancies [79] and suggesting a role for adjunctive vitamin D in LMWH-treated obstetric APS.