Introduction

Antiphospholipid syndrome (APS) is a complex acquired autoimmune disorder (isolated or coexistent with other autoimmune diseases) characterized by an increased risk for venous and arterial thromboembolism, associated with the presence of circulating antiphospholipid antibodies (aPL), including lupus anticoagulant (LA), immunoglobulin (Ig) G and IgM isotypes of anticardiolipin (aCL), and anti-β2 glycoprotein I (anti-β2GPI) antibodies.^1,2 Importantly, in the general population, the detection of these antibodies is also associated with a heightened thromboembolic risk.³

The traditional Sapporo (Sydney) classification criteria for APS are focused primarily on the presence of venous or arterial thrombosis or pregnancy morbidity in conjunction with persistent aPL positivity.⁴ This approach has been widely adopted in diagnosis of APS in everyday practice. In 2023, the updated American College of Rheumatology (ACR) and European Alliance of Associations for Rheumatology (EULAR) classification criteria for APS introduced a more refined point‑based scoring system with 6 clinical domains (macrovascular venous thromboembolism [VTE], macrovascular arterial thrombosis, microvascular thrombosis, obstetric, cardiac valve, and hematologic).⁵ According to the updated criteria, to classify a patient as having APS, a minimum of 3 points in the clinical and laboratory categories are required⁵; however, this approach results in major diagnostic and therapeutic challenges when managing patients previously diagnosed with APS.⁶

Atrial fibrillation (AF), the most common sustained cardiac arrhythmia of increasing prevalence, rises 5‑fold the risk of ischemic stroke (IS) and transient ischemic attack (TIA), and the risk persists despite treatment with direct oral anticoagulants (DOACs).⁷ We recently reported that up to 21% of patients with AF had positive aPL, including 15.2% with positive LA, 7.2% with positive aCL, and 7% with positive anti-β2GPI, and the presence of aPL was associated with increased stroke risk during follow‑up,⁸ suggesting that APS could partly explain the failure of DOAC therapy in these patients.

Notably, the current APS classification criteria may fail to adequately identify affected individuals among elderly patients or those with AF, due to atypical presentations and limited representation in earlier research. Therefore, this study aimed to evaluate the applicability and performance of the 2023 classification criteria for APS in an elderly AF population.

Patients and methods

Results

In the study group, 51 patients (21%) with AF tested positive for aPL, including 37 (15.2%) individuals with positive LA, 19 (7.8%) with moderate (n = 13; 68.4%) to high (n = 6; 31.6%) titer aCL, and 17 (7%) with moderate (n = 6; 35.3%) to high (n = 11; 64.7%) level of anti-β2GPI antibodies in both IgG and IgM isotypes. Persistent LA was detected in 9 patients (3.7%).

Based on the Sapporo criteria, we identified 33 patients (13.6%) with APS (Supplementary material, Table S1). Based on the 2023 updated APS classification criteria, we found that 26 patients (10.7%) scored 1 laboratory point, 4 (1.6%) scored 2 points, 1 (0.4%) 4 points, 11 (4.5%) 5 points, 6 (2.5%) 6 points, 2 (0.8%) 7 points, and 1 (0.4%) 8 points. The laboratory criterion for the 2023 APS classification was met by 21 patients (8.97%; Table 1). There were weak correlations of the 2023 APS laboratory score with ETP (R = 0.143; P = 0.01) and high‑density lipoprotein cholesterol levels (R = –0.132; P =0.03), but not with the other variables.

A total of 147 events (60.49% of patients) were recorded, including cerebrovascular ischemic events, systemic embolism (SE), and myocardial infarction. Of these, 117 events (79.59%) occurred within the 3 months preceding enrollment, and 30 events (20.41%) were observed during a median follow‑up period of 52 (IQR, 46.5–56) months. No cases of SE were observed during follow‑up. No past VTE events were reported, nor did any occur during follow‑up.

According to the 2023 APS clinical classification criteria, 5 patients (3.4%) received additional 2 points for thrombocytopenia, bringing their total score to 6 points. Additionally, 72 patients (48.98%) were assigned 4 points, while 70 patients (47.62%) received 2 points. The 2023 clinical classification criteria for APS were met by 77 patients (31.69%; Table 1).

Only 5 patients (2.06%) fulfilled both the clinical and laboratory classification criteria for 2023 APS within the 3‑year interval between aPL testing and the clinical event (Supplementary material, Table S1). Additionally, thrombocytopenia was observed in 14 patients (5.75%), of whom 3 (21.43%) were positive for aPL; however, this criterion did not affect the number of patients diagnosed using the 2023 APS criteria.

When comparing patients who met the laboratory classification criteria for 2023 APS with those who were aPL‑positive but did not fulfil the criteria, no differences were observed in baseline characteristics or laboratory test results, including coagulation parameters (Table 1).

Among the AF patients with thromboembolic events, those who met the 2023 APS clinical classification criteria had significantly fewer comorbidities than those who did not, including hypertension, diabetes mellitus, history of IS, and HF (Table 1). They also had 18.29% lower BMI, lower CHA₂DS₂-VASc scores, 18% lower blood glucose, 33.33% higher LDL cholesterol level, 15.68% lower fibrinogen level, and by 5.83% reduced TAFI antigen concentration (Table 1).

When comparing the AF patients who fulfilled the previous 2006 APS criteria (n = 12) but did not reach 3 laboratory points for the 2023 APS, with those who met the 2023 APS laboratory classification criteria (n = 21), no differences were observed between the 2 groups, except for a lower proportion of women (n = 2 vs n = 11; P = 0.049) and 27.16% longer CLT in the latter group (median [IQR], 103 [98.5–115] vs 81 [74–99] min; P = 0.02).

Discussion

Our findings demonstrate that, despite a high incidence of arterial thrombosis, affecting more than a half of the AF cohort, and a relatively high frequency of aPL positivity, as few as 2% of patients met both the laboratory and clinical 2023 ACR/EULAR APS classification criteria. This observation suggests that the current classification criteria may underestimate thrombotic risk in elderly populations with cardiovascular risk factors, and they are likely more suitable for young individuals, especially those with other autoimmune diseases.

Among patients with AF, new IS occurred at an annual rate of 4.7% (95% CI, 4.2%–5.3%) despite ongoing oral anticoagulant therapy,¹¹ yet these events are often attributed solely to AF, and aPL testing is rarely considered. In older AF patients, who often present with multiple comorbidities and a high inherent risk of arterial thrombosis, a single IS event is often insufficient to fulfil the 3‑point clinical requirement for the 2023 APS diagnosis, even in the presence of aPL positivity.⁵

Our study focused on AF‑related outcomes, and therefore did not assess obstetric complications, being, however, irrelevant due to the older age of female patients.

The aPL prevalence in the general population has been reported to be as high as 15.1%, with even higher rates observed in older individuals,¹² and it represents a major indication for lifelong anticoagulation if stroke or thrombosis has occurred.^12,13 However, data on aPL positivity specifically in AF patients remain limited. Tanne et al¹⁴ found intermediate or low aCL titers in 34% of AF patients with previous IS, while data from our cohort indicated that aPL positivity was observed in 21% of AF patients and was associated with an increased risk of IS despite DOAC treatment.⁸ These findings suggest that aPL may be more prevalent in AF patients than previously thought, and that its clinical manifestation as APS could also be more common than currently recognized. For this reason, the decision on the presence or absence of APS could be of key therapeutic relevance since warfarin is the preferred anticoagulant for most patients with thrombotic APS.¹⁵

In the current report, we found no differences in prothrombotic markers between the patients who met the 2023 APS laboratory classification criteria and those who were aPL‑positive but did not meet the criteria, suggesting that current laboratory definitions may offer limited added value beyond aPL positivity alone and may contribute to APS underestimation in studies focusing on patients with AF. Interestingly, the patients who met the previous 2006, but not the updated 2023 APS laboratory classification criteria, were characterized by 27% longer CLT, typically associated with a prothrombotic fibrin clot phenotype.¹⁶ However, K_s and the levels of major fibrinolysis inhibitors, such as PAI‑1 and TAFI, did not differ between the subgroups, indicating that altered fibrinolysis may be likely driven by other factors, for example, oxidative post‑translational modifications of fibrinogen (eg, oxidation).¹⁶

When comparing the patients with arterial thrombosis, those meeting the 2023 APS clinical classification criteria had significantly fewer comorbidities, supporting the notion that the current criteria tend to favor younger individuals free of comorbidities beyond autoimmune disorders. Additionally, the patients who met the 2023 APS clinical criteria, as compared with those who did not, showed lower levels of fibrinogen and TAFI, which are typically associated with increased cardiovascular risk,¹⁷ indicating that individuals from the former group were “healthier.” As a result of the new criteria, older or multimorbid patients with AF, who are generally at a high risk of arterial thrombosis, are less likely to score 3 points or more in the 2023 APS clinical criteria, even if their thrombotic risk is elevated.¹⁸

Among coagulation markers, only ETP, reflecting plasma potential for thrombin generation, correlated weakly with the APS laboratory score. Prior studies have shown that APS patients with a history of thrombosis tend to have higher ETP than healthy controls,¹⁹ though others have reported comparable ETP between APS patients and controls.²⁰ These findings suggest that ETP may have potential as a complementary marker for aPL‑related risk stratification, particularly in patients with AF, which requires further investigation.

This study has several limitations. First, due to the low number of patients meeting the 2023 APS classification criteria, reliable comparisons with those classified under the previous criteria were not feasible. As a result, we analyzed AF patients separately based on clinical and laboratory criteria to better assess their diagnostic value in this context.

Second, asymptomatic thromboembolic events, including venous thrombosis, may have been missed. Furthermore, we excluded patients with myocardial infarction or IS within 3 months prior to initial aPL testing to avoid potential false‑positive results. However, this exclusion may have further reduced the number of patients meeting the 2023 classification criteria for APS. Finally, larger studies are needed to assess the applicability of the new 2023 APS criteria in the AF population and to validate our findings. Nonetheless, our results suggest that some adjustments to the current classification may be warranted in this group.

In conclusion, the new approach to APS classification criteria from 2023 has several limitations, and further clinical research is needed to evaluate its applicability in elderly patients with CVDs. The present analysis demonstrates, at least in patients with AF, that the traditional 2006 APS criteria performed better and might help to identify individuals at a risk of thromboembolism while on DOACs, who might benefit from warfarin. In our opinion, the use of the Sapporo criteria should be preferred in complex older patients in whom APS as a contributor to thromboembolism is considered.