Discussion
In this mixed-methods study, we conducted a comprehensive characterisation of the clinical presentation of patients with aPL-associated chorea. Our findings provide valuable insights into clinical features, laboratory findings, radiological findings, treatments and outcomes of this rare manifestation. The key findings are: (1) Chorea was the initial symptom in 87.9% of patients, with most first consulting a neurologist. (2) Both the 2006 and 2023 APS classification criteria failed to identify more than half of the patients. (3) Contralateral striatal hypermetabolism on PET-CT/MRI is highly useful in diagnosing aPL-associated chorea. (4) GCs and ISTs are associated with a higher complete resolution rate.
Chorea can be due to a large number of aetiologies.99 While hereditary causes are the most common, identifying acquired or symptomatic chorea is crucial as these conditions are potentially treatable. Major causes of acquired chorea include infectious/postinfectious, pharmacological, vascular, metabolic and autoimmune conditions, such as autoimmune encephalitis, SLE and APS.100 Currently, there are no standard diagnostic criteria for APS-associated chorea. In this study, patients with a positive aPL test were included. Among them, only 41%–43% met the 2006 or 2023 APS criteria (table 1), significantly lower than the sensitivity reported in patients with general APS.5 This lower sensitivity may be due to the rarity of thrombotic events, especially venous thrombosis, and the younger age of many patients with chorea, precluding pregnancy morbidity. Although both are low, 2023 APS criteria are more sensitive than 2006 criteria in identifying APS-associated chorea. This is because the 2023 ACR/EULAR APS classification criteria introduced some non-criteria manifestations such as thrombocytopenia, microvascular diseases and valvulopathy.5
For many years, neuroimaging techniques have focused primarily on structural changes. Our data show that over half of patients with aPL-associated chorea had normal brain MRI results or non-specific white matter hyperintensities. Only 14.8% (20/135) exhibited basal ganglia lesions (table 2), which were often old and unrelated to the recent chorea, hence not providing additional diagnostic information. Recent advances in functional imaging have enabled the in vivo analysis of neuronal dysfunction and brain activity. PET is a functional imaging technique that involves the injection of a radiolabelled ligand. This ligand either binds to specific structures, such as neurotransmitter receptors, or integrates into the body’s tissues, as with 18F-fluorodeoxyglucose (18F-FDG). 18F-FDG PET imaging provides valuable information in cases of chorea with various aetiologies, offering insights into pathogenesis, disease course and basal ganglia functions. Striatal hypometabolism is associated with neurodegenerative causes of chorea, whereas striatal hypermetabolism tends to be seen in cases of chorea of transient and thus treatable aetiologies, including hyperthyroidism, Sydenham’s chorea and aPL-associated chorea.101 We identified 16 patients who had undergone PET-CT/MRI, 5 patients from PUMCH and 11 patients from the literature. Intriguingly, all of them showed contralateral striatal hypermetabolism, while none exhibited abnormalities on MRI (table 3). Moreover, contralateral striatal hypermetabolism is often observed as a functional correlate of chorea, with elevated metabolic activity detected in the striatum during chorea episodes. This heightened metabolism tends to normalise once the chorea subsides, indicating a reversible metabolic alteration associated with the condition (figure 1). This dynamic change underscores the importance of functional imaging in diagnosing and understanding the transient nature of chorea and its underlying pathophysiological mechanisms. Therefore, striatal hypermetabolism is not specific of aPL-associated chorea but one might consider including FDG-PET in the diagnostic workup of patients with chorea of unknown cause.
The pathogenesis of chorea in APS remains unclear, with three main hypotheses proposed.6 102 The first involves thrombosis or embolism formation without overt vascular changes. The second suggests autoimmune reactions targeting vascular endothelium, potentially inducing thrombosis formation or non-thrombotic vascular occlusion. The third hypothesis involves autoimmune reactions directly targeting the phospholipid-containing basal ganglia, causing neurotoxicity, loss of neuroplasticity and synaptic transition. Pathogenic autoantibodies may include aPL or unknown antineuron antibodies. In vitro data indicate that aPL can directly permeabilise and depolarise brain synaptoneurosomes (a composite particle containing one or more presynaptic compartments attached to a postsynaptic element).103 However, aPL titres typically remain stable regardless of the clinical course of chorea, prompting the search for new antibodies. Indeed, one study identified serum antistriatal antibodies in two patients, with titres dropping dramatically as chorea improved, while aCL levels remained unchanged during the disease course.62 Antineuron antibodies were also identified in patients with movement disorders associated with lupus and aPL by another study.73
Given the heterogeneity in chorea presentation and treatment responses, it is highly plausible that multiple mechanisms, rather than a single one, contribute to these abnormalities. Without biopsy availability, concrete answers remain elusive. However, one may speculate on the mechanisms based on clinical findings. For patients with bilateral involvement, subacute onset, normal MRI findings and a dramatic response to corticosteroids, autoimmune mechanisms targeting vascular endothelium or basal ganglia neurons likely play a crucial role. Conversely, for patients with unilateral involvement, acute onset, accompanying thrombotic complications (eg, cerebrovascular accidents or deep vein thrombosis), structural MRI abnormalities, and a response to aspirin or anticoagulants, thrombosis formation may be the primary mechanism.
There are no controlled prospective studies on the treatment of aPL-associated chorea. Current treatment regimens are based primarily on open-label studies and expert opinions. Neuroleptics were commonly used for symptomatic control of chorea, usually with good responses. While there have been cases where aspirin and/or anticoagulants led to complete chorea resolution, most patients received combination therapies, including aspirin, anticoagulants, steroids and immunosuppressants. Our analysis of treatment outcomes revealed that patients that received GC and/or IST had a higher complete resolution rate (61.5%) compared with those treated with AC and/or AP therapies alone (48.6%) (table 4). This suggests that for most patients, AC and/or AP alone may not be sufficient to completely eliminate chorea, and the addition of GC and IST may be necessary. Nevertheless, due to the variety of medication combinations used, it is not possible to infer definitive therapeutic guidelines.
Our study has several limitations. As a retrospective study, the data from medical records and particularly from the literature were incomplete, limiting the scope and robustness of further analyses and introducing a risk of reporting bias. The variability in aPL detection methodologies across different studies may account for differences in antibody results between our patients and those reported in the literature. We included only English and Chinese publications, which might result in the exclusion of relevant studies published in other languages, thereby missing potentially important data. Lastly, the literature tends to over-represent unusual, severe or positive outcomes, which could skew the reported clinical features and treatment effects.