The brain in antiphospholipid syndrome beyond thrombosis

1 Introduction

Antiphospholipid syndrome (APS) is a systemic autoimmune disorder defined by persistent antiphospholipid antibodies and clinical features that include vascular thrombosis and obstetric morbidity (1). Beyond these classification criteria, the central nervous system (CNS) is one of the most frequently and severely affected target systems (2). In clinical practice, the recognition of CNS involvement in APS has centered on ischemic stroke, especially in young patients or in cryptogenic stroke (3). Cognitive dysfunction and headache are common in APS, and less frequent manifestations such as seizures and chorea remain clinically important. Despite their clinical relevance, these manifestations are often not attributed to APS in routine clinical evaluation.

Cohort studies of APS report cognitive impairment in 11%–60% of patients, migraine in approximately 20%, seizures in approximately 7%, and chorea in 1.3% (4, 5). Collectively, these manifestations indicate that the neurological burden of APS extends well beyond ischemic stroke. Many patients report long-standing functional limitations from these manifestations, independent of any history of ischemic events. Antiphospholipid antibodies (aPL) act on the central nervous system through several non-thrombotic mechanisms, providing a coherent pathophysiological basis for the manifestations described above (2). These mechanisms include antibody-mediated neurotoxicity, microvascular injury, blood–brain barrier (BBB) disruption, and complement-driven neuroinflammation. Acting together within the neurovascular unit, they produce a continuum of injury that is not adequately captured by conventional MRI or the standard antibody panel.

This review focuses on the following four under-recognized non-thrombotic CNS manifestations of APS: cognitive dysfunction, movement disorders, headache, and seizures. These manifestations are considered expressions of a shared process termed neurovascular-immune dysfunction. A three-tier diagnostic framework is proposed to address them in clinical practice. Throughout this review, the evidence derives from heterogeneous populations, including primary APS, systemic lupus erythematosus (SLE)-associated APS, neuropsychiatric systemic lupus erythematosus (NPSLE), asymptomatic aPL carriers, and pediatric APS, and findings in one population do not necessarily generalize to the others, as indicated in the text and in Table 1.

Table 1. Evidence and treatment for CNS manifestations of APS beyond overt stroke.

Superscript numbers indicate references. Bracketed labels denote study design; bold marks findings extrapolated from NPSLE, aPL-carrier, or non-APS populations rather than APS-specific data. aPL, antiphospholipid antibodies; APS, antiphospholipid syndrome; DTI, diffusion tensor imaging; FDG-PET, fluorodeoxyglucose positron emission tomography; GABA-A, γ-aminobutyric acid type A; GFAP, glial fibrillary acidic protein; MRI, magnetic resonance imaging; NfL, neurofilament light chain; NPSLE, neuropsychiatric systemic lupus erythematosus; RCVS, reversible cerebral vasoconstriction syndrome; SLE-APS, systemic lupus erythematosus–associated APS; SPECT, single-photon emission computed tomography; SR, systematic review; TNF-α, tumor necrosis factor α.

2 Common ground across the four manifestations

Cognitive dysfunction, chorea, headache, and seizures are usually discussed separately in the APS literature. However, several clinical and mechanistic features are shared across the four, supporting a unified discussion.

Clinically, all four sometimes occur without demonstrable cerebral infarction. Conventional MRI is normal or shows only non-specific white matter changes in a sizable proportion of affected patients (4). The discrepancy between clinical severity and imaging findings often leads to misclassification as idiopathic, functional, or psychiatric disorders.

Mechanistically, each of the four manifestations has a credible non-thrombotic explanation, although the supporting evidence varies in strength and source, being strongest for cognitive dysfunction and chorea and weakest for seizures, where thromboembolic mechanisms remain the dominant established association. aPL bind to neuronal and glial surface targets (6), dampen γ-aminobutyric acid type A (GABA-A) receptor-mediated chloride currents (7), reduce endothelial nitric oxide bioavailability (8), and breach BBB integrity (9, 10). Animal models have reproduced several of these manifestations, including cognitive impairment and behavioral abnormalities, in the absence of overt infarction (9, 11, 12). Beyond the brain, aPL also drive platelet activation, complement deposition, and the release of neutrophil extracellular traps (NETs), processes that contribute to the broader thrombo-inflammatory phenotype of APS (13).

Therapeutically, the limits of anticoagulation alone are visible across all four manifestations. Hydroxychloroquine, immunomodulation, and symptom-directed treatment have produced responses in patients who are refractory to anticoagulation (4). These clinical, mechanistic, and therapeutic features support viewing the four manifestations as different expressions of a shared underlying process. This shared process is termed neurovascular-immune dysfunction (Figure 1). It integrates antibody-mediated neurotoxicity, microvascular injury, BBB disruption, and complement-driven neuroinflammation acting within the neurovascular unit.

The neurovascular unit is the functional assembly of endothelial cells, pericytes, astrocytic end feet, the basement membrane, neurons, and microglia. Under physiological conditions, it maintains blood–brain barrier integrity, couples regional cerebral blood flow to neuronal activity, and preserves an immunologically privileged microenvironment. In APS, aPL and associated thrombo-inflammatory mediators act on several of these components at once, so that endothelial activation, barrier breakdown, and glial inflammation reinforce one another rather than occurring in isolation. Response to immune-directed treatment provides a clinical test of this shared mechanism. Such response has been reported for cognitive dysfunction and chorea, whereas for headache and seizures, it remains untested, defining a gradient of evidence across the four manifestations.

3 Cognitive dysfunction without overt stroke

Cognitive dysfunction is the most common non-thrombotic CNS manifestation in APS and often the most disabling (14). Patients often report subjective cognitive symptoms, including “brain fog,” slowed thinking, reduced concentration, and difficulty performing everyday tasks (15). The symptoms fluctuate over time and are often too subtle to be captured by routine cognitive screening. Even mild impairment can affect employment, education, and social roles, and is associated with a sustained reduction in quality of life.

Cognitive impairment is reported in 11%–60% of patients with APS, with the variability reflecting differences in assessment methods, cohort selection, and the inclusion or exclusion of patients with prior stroke (4). In aPL-positive individuals without prior thrombotic or obstetric events, cognitive impairment has been reported at frequencies of 5.9%–31.1% in a systematic review (16). Cognitive impairment is more frequent and more severe in patients with lupus anticoagulant, higher aPL titers, or triple positivity. The risk of cognitive impairment is not uniform across antiphospholipid antibody profiles. Reported frequencies are higher in patients who meet the clinical criteria for APS than in asymptomatic aPL carriers. Lupus anticoagulant positivity, high antibody titers, and multiple positivity confer the greatest risk. Isolated low-titer or single-antibody positivity carries comparatively lower risk. The affected domains include executive dysfunction, complex attention, visual reproduction, and learning, supporting a profile of cognitive dysfunction that is not limited to global cognitive decline (4). In aPL-positive individuals, dementia is reported in 0%–6%, considerably less than the frequency of cognitive dysfunction (16).

Several non-thrombotic mechanisms have been implicated in APS-related cognitive impairment (2). Loss of BBB integrity allows aPL and complement components to reach brain tissue, where they can contribute to local inflammation. aPL can target neuronal and glial antigens, and passive transfer of patient-derived antibodies produces cognitive impairment in animal models (11). In experimental APS (eAPS), β2GPI immunization elevates brain tumor necrosis factor α and prostaglandin E (PGE) with microglial reactivity in the cortex, parahippocampal region, and hippocampus. These changes are accompanied by behavioral hyperactivity (17). In factor V Leiden eAPS mice, β2GPI immunization produces demyelination and axonal loss with impaired learning, attributed to antibody-mediated neurodegeneration (18).

A standard workup with conventional MRI and brief cognitive screens, such as Mini-Mental State Examination (MMSE) or Montreal Cognitive Assessment (MoCA), frequently shows normal or non-specific findings in patients with APS and documented cognitive complaints (14). Diffusion tensor imaging (DTI) can reveal microstructural white matter abnormalities in patients with APS whose conventional MRI is unremarkable, including individuals without overt neurological symptoms (19). Brain single-photon emission computed tomography (SPECT) shows perfusion defects in patients with APS that correlate with cognitive impairment, even when conventional MRI is normal (20). Immune-directed treatment has been examined in APS. In an open-label pilot trial of rituximab in aPL-positive patients with non-criteria manifestations, cognitive dysfunction improved in four of five treated patients, without a consistent change in aPL titers (21).

4 Chorea

Chorea is the most characteristic movement disorder of APS and, despite occurring in only 1.3% of patients, can be the initial manifestation of the disease (4, 5). The movement is typically choreiform, unilateral or bilateral, and develops over days to weeks, with spontaneous remission occurring in some patients. Recurrence is not uncommon; in the largest series of APS-related chorea, 17 of 50 patients had more than one episode (22). The presence of chorea in a young patient, particularly when accompanied by other features such as livedo, cytopenias, or pregnancy loss, should prompt evaluation for antiphospholipid antibodies.

Chorea predominantly affects young women, with a mean age at onset of 21 ± 12 years in the largest APS-chorea series (22). Chorea shows a particularly strong association with childhood-onset APS, occurring in 14% of patients with disease onset before age 15 compared with 1% of those with later onset (5). Hormonal factors appear to contribute, with oral contraceptive use preceding chorea onset in 12% of cases and pregnancy or the postpartum period in 8% (22). Together, the demographic profile of APS-related chorea, characterized by young women, frequent childhood onset, and hormonal triggers, points to mechanisms beyond classical thrombosis.

Several lines of evidence indicate that APS-related chorea is not adequately explained by ischemic mechanisms. Conventional brain imaging is often normal or non-specific despite overt chorea, and basal ganglia lesions are rarely demonstrated (22). Fluorodeoxyglucose positron emission tomography (FDG-PET) reveals striatal hypermetabolism in APS-related chorea, suggesting active neural dysfunction rather than tissue loss (23). Consistent with this functional disturbance, antibodies from pediatric patients with APS-related chorea bind to a dopaminergic neuronal cell line, a proof-of-principle finding that is consistent with a direct antibody–neuron interaction (24). This evidence supports an antibody-mediated pathogenesis and suggests that immunomodulation may complement anticoagulation in clinical management.

An antibody-mediated pathogenesis has practical consequences for how APS-related chorea is diagnosed and treated. Diagnostic evaluation should include standardized clinical assessment with video documentation, validated movement scales, and a careful history of hormonal triggers (25). Dopamine antagonists, particularly haloperidol, provide effective symptomatic control of involuntary movements. Consistent with an antibody-mediated mechanism, glucocorticoids are the most frequently reported treatment for refractory cases, and immunomodulatory agents, including intravenous immunoglobulin, plasma exchange, and cyclophosphamide, have been used (25). Rituximab has also been reported to be effective in therapy-resistant childhood APS-related chorea (26). Because this evidence derives from case reports and small case series, immunomodulation should be individualized rather than regarded as a standard recommendation. Other movement disorders, including Parkinsonism, dystonia, and paroxysmal dyskinesia, have been described, almost always as isolated cases and frequently refractory to conventional therapy (4, 27). Their rarity and the absence of controlled data mean that a causal link with APS remains uncertain.

5 Headache/migraine

Headache is among the most frequent non-thrombotic CNS manifestations of APS and is often attributed to a primary headache disorder rather than recognized as a feature of the syndrome (4, 5, 28). The associated burden on quality of life is substantial and often precedes the diagnosis of APS by several years. The new onset of headache or a change in its established pattern in a young patient should prompt evaluation for antiphospholipid antibodies, particularly when other clinical features of the syndrome are present. Migraine is the most frequent type, reported in approximately 20% of patients with APS in the largest cohort to date (5). Migraine occurs approximately twice as frequently in female patients as in male patients with APS (23% vs. 12%). Onset is typically in young adulthood, with the majority of cases presenting between the second and fourth decades of life (29). This demographic profile overlaps substantially with that of primary migraine, and distinguishing the two requires attention to associated features of the syndrome rather than to headache characteristics alone.

APS-related headache arises largely through non-thrombotic mechanisms. aPL induce endothelial activation in cerebral vessels, promoting vasoconstriction through mediators such as endothelin-1 (4, 30). aPL also alter platelet activation and serotonin metabolism, processes that have been linked to the pathogenesis of migraine independent of APS (28). These vascular and immunological processes converge on the cerebral microvasculature and provide a mechanistic basis for headache that develops without overt cerebral infarction. The non-thrombotic pathogenesis of APS-related headache has direct implications for both clinical recognition and treatment. Clinical recognition should include screening for aPL in young patients with migraine that is refractory to standard therapy, accompanied by aura, or associated with other features of the syndrome (29). Conventional migraine-specific therapy provides symptomatic control, and antithrombotic treatment of the underlying syndrome has produced substantial improvement in patients with refractory headache, an observation that requires prospective confirmation and does not justify initiating antithrombotic therapy for headache alone (31). Reversible cerebral vasoconstriction syndrome (RCVS) has been reported in association with APS and should be considered in patients presenting with thunderclap headache (32, 33). Rarely, APS-related headache reflects a CNS vasculitis phenotype with multifocal cerebral arterial stenosis, in which immune-directed treatment has been reported to improve the arterial stenosis (34).

6 Seizures without ischemic lesions

Seizures are an established neurological manifestation of APS and occur more frequently in patients with associated SLE than in those with primary disease. Both generalized tonic-clonic and focal seizures occur in APS. A new seizure of undetermined cause in a young patient warrants testing for antiphospholipid antibodies (35). Seizures occur in approximately 7% of patients with APS, a frequency nearly 10 times that of the general population (5, 35). Cerebral thromboembolic events are the factor most strongly associated with seizures, although seizures also occur without a demonstrable infarct (36). Poststroke epilepsy reaches 17% in APS, higher than the rate of approximately 10% in the general poststroke population.

The mechanisms underlying APS-related seizures involve both microthrombosis and a direct immune-mediated effect on neurons (35, 36). Vascular injury in APS is both arterial and venous. Arterial ischemic stroke produces cortical infarcts that can serve as epileptogenic foci. Cerebral venous thrombosis also causes parenchymal injury, including venous infarction with a frequent hemorrhagic component, and provides a similar substrate for seizures (37). Acute symptomatic seizures occur in 30%–40% of patients with cerebral venous thrombosis, a higher proportion than in arterial stroke (37). Small-vessel thrombosis contributes additional cortical microinfarcts. In eAPS, aPL accumulate in inhibitory cortical and hippocampal neurons, and patient-derived antibodies reduce inhibitory GABA-A receptor signaling (7, 18).

This dual pathogenesis determines the diagnostic and therapeutic approach to APS-related seizures. Conventional MRI lacks a specific pattern for APS-related seizures and is normal in 30%–40% of patients (35). Antiphospholipid antibodies merit testing in young patients with atypical seizures, particularly alongside thrombocytopenia, livedo reticularis, or valvular disease. Management combines antiepileptic treatment with anticoagulation when there is evidence of a thrombotic event. The immune contribution to seizures is supported mechanistically rather than therapeutically. Whether an anticoagulation-refractory, immune-driven subgroup exists that might respond to immune-directed treatment is an open question for future study.

7 A unified diagnostic framework

This review proposes a three-tier framework for the layered evaluation and clinical attribution of these neurological manifestations. The first tier is clinical recognition and high-resolution phenotyping, the second is structural and functional imaging that captures what conventional MRI misses, and the third is mechanistic biomarkers. This framework is illustrated in Figure 2, and its application to the four manifestations is summarized in Table 1. This framework is conceptual and has not been validated.

Underdiagnosis stems less from rarity than from a standard workup poorly suited to these presentations. Clinical disease activity plotted against the results of that workup defines four scenarios (Figure 2). Three correspond to familiar APS phenotypes. The fourth, active disease with normal findings, is the diagnostic gap, where the four manifestations discussed here fall, clinically active yet radiologically silent (4, 14, 25, 35).

Tier 1 begins with a low threshold for suspicion. A comprehensive aPL profile, including non-criteria antibodies where available, is most useful when obtained early rather than after other causes are exhausted (4).

Tier 2 selects imaging by symptom. In cognitive dysfunction, DTI detects microstructural white-matter injury when conventional MRI is unremarkable (19), and SPECT reveals perfusion defects that correlate with cognitive impairment (20). In chorea, FDG-PET shows striatal hypermetabolism, which indicates active dysfunction rather than tissue loss (23). In migraine, magnetic resonance angiography and vessel-wall imaging are indicated when RCVS is suspected (32, 33). In seizures, MRI is non-specific or normal in a large proportion of patients (35), so structural imaging adds little, and the evaluation depends on phenotyping and mechanistic markers.

Tier 3 links these findings to mechanism. Serum neurofilament light chain and glial fibrillary acidic protein (GFAP) quantify neuronal and astrocytic injury and perform well in active major NPSLE, though their value in APS without SLE is undefined (38). Non-criteria aPL improve sensitivity in seronegative APS and refine risk stratification. Markers of complement activation reflect the thrombo-inflammatory axis and may identify candidates for complement-directed therapy (13).

The threshold for moving beyond Tier 1 should fall in patients with high-risk antibody profiles or concurrent SLE (4). A high-risk antibody profile, in particular triple positivity, carries the strongest association with thrombotic and overall disease burden. In these patients, advanced imaging and mechanistic biomarkers warrant earlier use rather than reservation for refractory cases.

8 Future directions and conclusion

Three areas remain open for future study. The first is the thrombo-inflammatory axis; complement activation, neutrophil extracellular traps, and platelet activation define candidate pathways and therapeutic targets. The second is antibody-mediated CNS injury; the binding of aPL to neuronal and glial targets needs characterization in APS specifically, as does the validation of injury markers, such as neurofilament light chain and GFAP, outside overt SLE. The third is the boundary between APS and SLE, whose shared and distinct CNS features remain incompletely defined.

The CNS involvement of APS extends beyond ischemic stroke. Cognitive dysfunction, chorea, headache, and seizures recur as non-thrombotic manifestations with a common basis in neurovascular-immune dysfunction. Understanding them in these terms and applying a structured diagnostic framework may improve the care of patients whose neurological symptoms are not explained by conventional investigation.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. The author verifies and takes full responsibility for the use of generative AI in the preparation of this manuscript. Generative AI was used as follows: Google Gemini for the schematic illustrations (Figures 1, 2), and Anthropic Claude for language editing of the text. All AI-assisted content was reviewed and verified by the author.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Zaida G. Ramirez-Oritz, UMass Memorial Health Care, United States

Reviewed by: Juexian Song, Capital Medical University, China

Ecem Sevim, National Institutes of Health (NIH), United States