This was a single-center retrospective cohort study performed at the Department of Rheumatology, Peking University People's Hospital. A total of 332 patients with APS were consecutively enrolled between July 2009 and January 2021. The diagnosis of APS was confirmed by two rheumatologists (YZG and YWZ) according to the 2006 Sapporo criteria (8) or catastrophic antiphospholipid syndrome (CAPS) according to the current diagnostic criteria (9). Another inclusion criterion was the disease onset age, which was >18 years. The study was approved by the ethics committee of Peking University People's Hospital (2019PHB253-01) and complied with the Declaration of Helsinki guidelines.

Baseline demographics and clinical and laboratory characteristics were obtained from the electronic medical records at the time of APS diagnosis. Venous thromboembolic events (e.g., deep venous thrombosis of the upper limbs of the legs, visceral venous thrombosis, and/or pulmonary embolism) were confirmed by limb ultrasound, pulmonary computed tomography (CT) or scintigraphy (ventilation/perfusion), abdominal pelvic CT scan, and vessel angiography as indicated. Arterial thrombotic events (e.g., peripheral arterial thrombosis, acute cerebral infarction, and/or visceral arterial thrombosis) were diagnosed using typical clinical pictures with positive arteriography [e.g., leg or upper limb ultrasound, CT, or magnetic resonance angiography (MRA)] and surgery. Multiple organ thrombosis was defined as the involvement of at least two organ systems during the course of the disease. The adjusted global antiphospholipid syndrome score (aGAPSS) was calculated for each patient by adding the points corresponding to the risk factors, excluding antibodies to phosphatidylserine/prothrombin (aPS/PT) that are not routinely tested in most clinical laboratories, as previously described (10). The aGAPSS ranged from 0 to 17.

Atherosclerosis was defined as the presence of any plaque in the carotid or femoral arteries, which were tested using B-mode and Doppler ultrasound examination with an Aplio 500 system (Canon Medical Systems Corp., Tochigi, Japan) with a 14L5 transducer or a Logic E9 system (GE Healthcare, Chicago, IL, USA) with a 9L transducer. Electrocardiography (ECG) was performed using a MAC 5500 HD resting ECG system (GE Healthcare, Wauwatosa, USA). Transthoracic echocardiogram (TTE) evaluation was performed using either an iE33 (S5-1 probe, Philips Medical Systems, Andover, Massachusetts, USA) or a Vivid E9 (GE Vingmed, Horten, Norway, UK) scanner with a 2.5–3.5 MHz transducer. AMI was diagnosed according to the fourth universal definition of myocardial infarction and classified as ST segment elevation myocardial infarction (STEMI), non-ST segment elevation myocardial infarction (NSTEMI), and unstable angina (UA) (11). Myocardial infarction with non-obstructive coronary arteries (MINOCA) was defined as no coronary artery stenosis > 50% in any potential infarct-related artery without other clinically overt specific causes for the acute presentation (12).

Patients were followed up for 5 years or monitored up to 31 December 2021 if the patients were enrolled after 30 December 2016 in outpatient services. Follow-up information was also obtained from electronic medical records or regular medical examination reports. Medication data were recorded, including sustained anticoagulation treatment, antiplatelet therapies, corticosteroids, immunosuppressants, and statins. If patients received warfarin, the international normalized ratio (INR) was documented every 3 months and the mean INR was calculated. The primary clinical outcome was all-cause mortality (defined as the time from recruitment to death from any cause). The second clinical outcome was the recurrence of thrombosis. Thrombotic events were independently adjudicated by two investigators (YZG and YWZ).

Continuous data with normal distribution were expressed as the mean ± standard deviation, and differences between groups were analyzed using one-way ANOVA. Continuous data with skewed distribution were expressed as medians (P25, P75), and differences between groups were analyzed using the Kruskal–Wallis test. Dichotomous variables were reported as frequencies (percentages), and differences between groups were compared using the chi-square test (or Fisher exact test when appropriate). Univariate and multivariate logistic regression analyses with an imputed dataset were adopted to identify risk factors for AMI, adjusted for age and sex. The variables assessed in the group differences were entered as independent variables in the univariate logistic regression analysis when the p < 0.1. The variables assessed in the univariate regression analysis were entered as independent variables in the multivariate logistic regression analysis when the p < 0.1. Survival and recurrence of thrombosis were estimated using the Kaplan–Meier method, and differences were evaluated using a stratified log-rank test. Data analyses were performed using SPSS 23.0 for Windows. Two-sided p < 0.05 was considered statistically significant.

Of the 332 APS patients, 93 had isolated obstetric APS, 207 had isolated thrombotic APS, and 32 had thrombotic APS with obstetric complications. A flow diagram of the individuals at each stage is shown in Figure 1. A total of 239 patients with thrombotic APS (207 isolated thrombotic and 32 thrombotic APS with obstetrical complications) were enrolled in our cohort. Follow-up data were available for 196 patients (82.0%) with an overall median follow-up time of 4.5 years. Of the 196 patients, 102 (52.0%) completed a 5-year follow-up and 21 (10.7%) died within 5 years.

The incidence of AMI was 12.1% (29/239) and detailed clinical profiles are shown in Table 1. Nine patients with AMI were male, and the average age of AMI onset was 44.6 years (Table 2). Of the 29 patients with AMI, 14 (48.3%) had STEMI, 5 (17.2%) had NSTEMI, and 10 (34.5%) had UA. In total, 13 patients with APS (44.8%) developed AMI before APS confirmation (range, 7 months to 30 years), 9 patients (31.0%) developed AMI after APS diagnosis (range, 1 month to 5 years), and 7 patients (24.2%) presented with AMI simultaneously with APS (confirmations were performed after 12 weeks). In total, 1 patient (3.4%) underwent coronary artery bypass grafting, 1 patient (3.4%) underwent thrombolysis, 9 patients (31.0%) underwent percutaneous coronary intervention, and the remaining 18 patients (62.2%) received conservative medical treatment. Of the 22 patients who underwent coronary angiography, 7 (31.8%) were MINOCA, including UA, STEMI, and NSTEMI (5, 2, and 1, respectively). One of the patients with AMI died within the first 30 days.

Next, we compared baseline demographics, clinical characteristics, and laboratory parameters between patients with and without AMI (Table 2). Patients with AMI had multiple organ thrombosis (55.1 vs. 34.3%, p = 0.029), recurrent thrombosis (58.6 vs. 34.3%, p = 0.011), higher incidence of atherosclerosis (62.1 vs. 23.8%, p < 0.001), higher neutrophil count (×10⁹/L) [4.68 (3.25, 8.17) vs. 3.71 (2.64, 5.80), p = 0.036], and fewer venous thrombosis events (27.6 vs. 63.3%, p < 0.001). Demographic data, other clinical or laboratory features, and aGAPSS scores did not show any significant differences between the two groups. Supplementary Table 1 showed the detailed profiles of accompanied autoimmune diseases.

For cardiac manifestations, the QTc interval (ms) was significantly longer in APS patients with AMI [438 (423, 454) vs. 425 (410, 446), p = 0.016]. However, significant differences were not observed in other cardiac features, including arrhythmia, myocardial hypertrophy, echocardiographic parameters, and percentage of hydroxychloroquine usage (shown in Table 3).

In logistic regression analysis, the variables assessed in the previous analysis were entered as independent variables with a cut-off p < 0.1. Results of the univariate and multivariate logistic analyses are shown in Table 4. In univariate logistic analysis, several variables were positively associated with AMI, including multiple organ thrombosis [odds ratio (OR) 2.715, 95% confidence interval (CI): 1.230–5.995, p = 0.013], recurrent thrombosis (OR 2.359, 95%CI: 1.076–5.174, p = 0.032), atherosclerosis (OR 5.236, 95%CI: 2.319–11.82, p < 0.001), and elevated neutrophil count (>6.3 ×10⁹/L) (OR 3.000, 95%CI: 1.328–6.780, p = 0.008). Interestingly, venous thrombosis was negatively associated with AMI (OR 0.221, 95%CI: 0.093–0.522, p = 0.001). With age and sex adjustment, the multivariate logistic models revealed that multiple organ thrombosis (OR 8.862, 95%CI: 1.817–43.212, p = 0.007), atherosclerosis (OR 5.397, 95%CI: 1.943–14.994, p = 0.001), and elevated neutrophil count (>6.3 ×10⁹/L) (OR 3.271, 95%CI: 1.268–8.440, p = 0.014) might be risk factors for AMI, and venous thrombosis (OR 0.106, 95%CI: 0.036–0314, p < 0.001) might be protective factors for AMI. Traditional CVD risk factors (smoking, diabetes mellitus, hypertension, dyslipidemia, chronic kidney disease, and hyperuricemia) were not associated with AMI in APS.

The median follow-up duration was 4.8 years in the AMI group and 4.5 years in the non-AMI group. At the last follow-up visit, 4/27 patients in the AMI group and 17/167 in the non-AMI group died, and overall survival rates between the two groups were not significantly different (Figure 2A). Detailed causes of death in these 21 patients with APS were shown in Table 5. In the AMI group, two patients died of pulmonary embolism and the other two patients died of severe infection. In the non-AMI group, the patients died of multiple causes, including pulmonary embolism (n = 5), severe infection (n = 2), malignant tumors (n = 2), severe thrombocytopenia (n = 1), disseminated intravascular coagulation (n = 1), and macrophage activation syndrome (n = 1).

As shown in Table 6, APS patients with AMI received more antiplatelet treatments [17 (63.0%) vs. 44 (26.3%), p < 0.001] and fewer immunosuppressants [10 (37.0%) vs. 123 (73.6%), p < 0.001] compared to APS patients without AMI. The application of anticoagulants, hydroxychloroquine, and corticosteroids was not significantly different between the two groups. The mean INR of APS patients who took the vitamin K antagonist was not significantly different between the AMI group and non-AMI group (2.19 ± 0.89 vs. 2.24 ± 0.68, P = 0.766). During the 5-year follow-up, the recurrence rate of overall thrombosis was 18.6% (5/27) in the AMI group and 13.7% (23/167) in the non-AMI group; recurrence of venous thrombosis was 3.7% (1/27) in the AMI group and 11.3% (19/167) in the non-AMI group, and the recurrence of arterial thrombosis was 14.8% (4/27) in the AMI group and 4.8% (8/167) in the non-AMI group, respectively. Kaplan–Meier analysis showed that there was no significant difference between the AMI group and the non-AMI group in recurrent rates of overall thrombosis (Figure 2B) or venous thrombosis (Figure 2C). However, the recurrence rates of arterial thrombosis differed significantly between the two groups (Figure 2D).

This study has several limitations. First, due to the relatively small number of patients with AMI, we did not compare the clinical and laboratory differences among UA, STEMI, and NSTEMI, and the effectiveness of different therapies in the acute phase of AMI, despite the lower percentage of invasive treatment. Second, because of the limitations of this retrospective study, we did not discuss the monitoring function of neutrophils or the therapeutic role of neutrophil-targeting therapy in AMI. Therefore, prospective cohort or multicenter studies are needed in the future.