Population

Between January 2014 and March 2017, a total of 292 participants with pregnancy‑related complications were referred to the Department of Thromboembolic Disorders, Institute of Cardiology in Kraków, Poland, with a suspicion of APS. All participants fulfilled at least 1 of the following clinical obstetric criteria: 1) at least 3 consecutive prefetal (before 10 weeks of gestation) and / or early fetal (from 10 weeks + 0 days to 15 weeks + 6 days) deaths, 2) fetal death (from 16 weeks + 0 days to 33 weeks + 6 days) in the absence of either PEC or PI with severe features, 3) either PEC or PI with severe features (<⁠34 weeks + 0 days) with / without fetal death, or 4) both PEC and PI with severe features (<⁠34 weeks + 0 days) with / without fetal death. PEC and PI were defined according to standard definitions.¹¹ All participants were enrolled in the study at least 3 months after the last obstetric complication. The control group was selected from women free of OAPS, matched for age, body mass index (BMI), and type of obstetric complications, who were diagnosed in our center during the same time period.

Exclusion criteria were as follows: prior TE (including documented pulmonary embolism [PE], deep vein thrombosis [DVT], ischemic stroke, peripheral artery embolism, or myocardial infarction), exacerbation of systemic lupus erythematosus (SLE), C‑reactive protein (CRP) level above 10 mg/l, signs of acute infection, known cancer, use of hormonal contraceptives, pregnancy, kidney failure (serum creatinine ≥120 µmol/l), severe liver injury, anemia (hemoglobin <⁠10 g/dl) inherited thrombophilia (ie, protein C deficiency, protein S deficiency, antithrombin deficiency, factor V Leiden and prothrombin G20210A mutation), and current anticoagulation. The study flowchart is presented in Supplementary material, Figure S1.

We collected demographic data and information on risk factors along with comorbidities. SLE was diagnosed based on the American College of Rheumatology classification criteria.¹² Arterial hypertension was defined as systolic blood pressure equal to or above 140 mm Hg and / or diastolic blood pressure equal to or above 90 mm Hg or taking antihypertensive medications. Obesity was defined as BMI equal to or above 30 kg/m². Diabetes mellitus was defined in accordance with the American Diabetic Association criteria.¹³ Smoking status was positive if a patient smoked at least 1 cigarette a day.

The Jagiellonian University Medical College Ethics Committee approved the study (1072.6120.316.2022), and the participants provided their written informed consent in accordance with the Declaration of Helsinki.

Antiphospholipid syndrome diagnosis

OAPS was diagnosed based on the criteria comprising at least 3 points from the clinical domain and at least 3 points from the laboratory domain.⁶ All aPL were determined at least 3 months after the end of pregnancy / the end of a pregnancy‑related complication, and all aPL measurements were performed twice if positive, at least 12 weeks apart to exclude clinically unimportant or transient changes in antibody levels. LA was assessed as recommended.¹⁴

The levels of immunoglobulin (Ig) G and IgM isotypes of aCL, and aβ2GPI along with antiphosphatidylserine / prothrombin (aPS/PT) antibodies were measured with the enzyme‑linked immunosorbent assay (ELISA) kits (INOVA, San Diego, California, United States). The division into moderate (40–79 IU) and high (≥80 IU) levels of aCL/aβ2GPI IgM/IgG was based on ELISAs.⁶ The cutoff values for both IgG and IgM isotypes of aPS/PT antibodies, that is, above 30 IU, were suggested as positive by the manufacturer.

The women with OAPS were divided into the high‑risk and low‑risk aPL profile groups according to the European League Against Rheumatism (EULAR) recommendations.⁶ The high‑risk aPL profile was defined as the presence of LA (on at least 2 occasions at least 12 weeks apart), or of double (any combination of LA, aCL, and / or aβ2GPI antibodies at moderate / high levels) or triple (all 3 subtypes) aPL positivity, or the presence of persistently high aPL titers. The women with the low‑risk aPL profile presented isolated aCL or aβ2GPI antibodies at low‑to‑medium titers.

Follow‑up

Follow‑up was started on the day when preliminary laboratory evaluations to confirm eligibility were performed, and was carried out every 6 months with visits to the center or phone calls. Clinical events recorded during follow‑up were TEs. A TE was defined as documented venous or arterial thrombosis.¹⁵ Venous thromboembolism (VTE) was confirmed after diagnosis of either DVT, established based on clinical symptoms and a positive finding on color duplex sonography, or PE, identified based on clinical symptoms and positive results of computed tomography angiography. Stroke was diagnosed on the basis of a new and abrupt neurological deficit lasting for at least 24 hours and typical lesions on brain imaging. Myocardial infarction was diagnosed based on typical chest pain and electrocardiographic changes in at least 2 contiguous leads and elevated cardiac troponin levels.

In the absence of a TE, the last visit at the center or telephone contact before June 2023 was considered the end of follow‑up.

Laboratory investigations

Venous blood samples were drawn with minimal stasis using atraumatic venipuncture between 8.00 AM and 10.00 AM, after an overnight fast and a 10‑minute rest.

CRP, fibrinogen, activated partial thromboplastin time (APTT), prothrombin time / international normalized ratio, and D‑dimer were determined using routine laboratory techniques.

Von Willebrand factor activity (ristocetin cofactor activity) was determined turbidimetrically (BCS XP System, Siemens, Erlangen, Germany), and the reference range was 58%–172%. PAI‑1 antigen was determined by ELISA (American Diagnostica, Stamford, Connecticut, United States). Thrombin activatable fibrinolysis inhibitor (TAFI) antigen was determined with an ELISA (Chromogenix, Lexington, Massachusetts, United States). Antiplasmin and plasminogen levels were measured in plasma with chromogenic assays (STA Stachrom 2‑antiplasmin and STA Stachrom plasminogen, Diagnostica Stago, Asnieres, France), and the reference range was 75%–135%.^3,16,17 Thrombophilia screening, which comprised protein C, free protein S, antithrombin, factor V Leiden, and prothrombin G20210A genetic variants, was performed in all study participants, as described before.¹⁸

Statistical analysis

The study was powered to have a 90% chance of detecting at least 10% difference in K_s and CLT using a P value of 0.05 based on the values of K_s and CLT established in our laboratory over the last 10 years for healthy individuals (n = 100; mean [SD] K_s, 7.02 [1.09] × 10^-9 cm² and mean CLT, 91 [16] min) and APS patients (n = 100; mean K_s, 6.39 [0.69] × 10^-9 cm² and mean CLT, 106 [21] min). To demonstrate such a difference or greater in K_s, at least 39 participants were required per group, while for CLT, at least 29 were needed.

Categorical variables are presented as numbers and percentages. Continuous variables are expressed as mean (SD) or median with interquartile range (IQR), as appropriate. The Shapiro–Wilk test was used to assess conformity with a normal distribution, whereas non‑normally distributed data were analyzed with the Mann–Whitney test. Categorical variables were analyzed using either the χ² test or the Fisher exact test. The Spearman rank correlation coefficient (R) was calculated to assess linear associations between the variables. The associations between the variables were examined using logistic regression and expressed as odds ratios (ORs) with 95% CI. The Kaplan–Meier analysis was used to estimate the event‑free rate in the study group, and the results were presented as hazard ratio with 95% CI. Two‑sided P values below 0.05 were considered significant. The statistical analysis was performed with the STATISTICA 13.0 software (StatSoft, Kraków, Poland).

Comparison of the patients with obstetric antiphospholipid syndrome and controls at baseline

The final analysis included 62 women with OAPS and 62 well‑matched controls (Table 1). In both groups, the most common obstetric complication was PEC or PI with severe features (<⁠34 w + 0 d) with / without fetal death (80.6% in the OAPS group and 79% in the controls). The mean (SD) time from the last obstetric complication to the study enrollment was 11.5 (4.6) months. SLE and other autoimmune diseases were observed only in the OAPS patients with at least 1 fetal loss from the 10th week of pregnancy (Supplementary material, Table S1).

Twenty‑one (33.9%) patients with OAPS were positive for LA, 58 (93.5%) for aCL, and 51 (82.3%) for aβ2GPI, while 38 (61.3%) had positive aPS/PT antibodies, all of them detected after a median (IQR) of 10 (7.3–13) months since the last obstetric complication. Nineteen OAPS patients (30.6%) had a triple‑positive aPL profile (LA + aCL + aβ2GPI).

There were no intergroup differences regarding the clinical and laboratory variables, except for by 10.5% higher CRP level in the OAPS group (Table 1). The patients with OAPS had slightly higher thrombin generation capacity, evidenced by ETP (Table 1) that positively correlated with aβ2GPI IgG levels (R = 0.39; P = 0.02), but not with other aPL or APS types.

An analysis of plasma fibrin clot properties (Table 1) showed that the patients with OAPS had lower K_s than the controls (by 16.4%; Figure 1A), indicating more compact fibrin networks (Supplementary material, Figure S2), and prolonged CLT (by 16.9%; Figure 1B; both P <⁠0.001). K_s negatively correlated with fibrinogen level (R = –0.43; P = 0.001). In the OAPS group, K_s negatively correlated with PAI‑1 level, as well as with aCL IgG, aβ2GPI IgG, and aPS/PT IgG (Supplementary material, Table S2). CLT was associated with PAI‑1 level, but not with TAFI, and also with aPS/PT IgG levels (R = 0.29; P = 0.02), but not with other aPL (Supplementary material, Table S2).

We observed lower, though within the reference range, plasminogen (by 7.9%) and higher TAFI antigen (by 15.6%) levels in the OAPS patients than in the controls, without intergroup differences in PAI‑1 or antiplasmin levels (Table 1).

No differences in fibrin clot properties and fibrinolysis proteins were reported regarding the type of obstetric complications (Supplementary material, Table S1).

Comparison of different antiphospholipid antibody profiles

High‑risk aPL OAPS participants (n = 47; 75.8%) had by 10.4% longer APTT and by 4.2% higher ETP than the remaining OAPS patients. The participants with a high‑risk vs low‑risk aPL profile had by 14.5% decreased K_s, by 12.2% longer CLT, and by 15.2% higher PAI‑1 antigen, without any other differences in the variables tested (Table 1). The patients with positive LA (alone or in combination with other aPL; n = 47; 75.8%), had by 16.8% lower K_s (Supplementary material, Figure S3A) and by 10.8% prolonged CLT (Supplementary material, Figure S3B) than those with LA‑negative OAPS.

There were no differences between the single‑positive aPL patients and the controls in laboratory paremeters related to coagulation and fibrinolysis as well as and fibrin clot properties (Supplementary material, Table S3).

Thromboembolic events

None of the patients had a TE before the study enrollment. None of the participants were lost to follow‑up (median [IQR] duration, 54 [49–58] mo). We recorded 10 (16.1%; 2.2/y) TEs in 10 participants with OAPS, including 9 from the high‑risk aPL group, and 3 (4.8%) TEs in the control group (Table 1) during follow‑up. Eight patients with OAPS experienced VTE, including 6 with DVT and 2 with PE, whereas 2 individuals experienced ischemic stroke. In the control group, 2 individuals experienced DVT and 1 woman had ischemic stroke. The participants with TE during follow‑up more commonly had hypercholesterolemia and autoimmune disorders as well as positive LA than the participants free of TE (Table 2).

Higher ETP (by 4.5%) was observed in the patients with OAPS and TE, as compared with the remaining patients (Table 2). Interestingly, the individuals with OAPS who experienced TE had by 17.5% lower K_s (Table 2; Figure 2A) and by 12.2% longer CLT than those without TE (Table 2; Figure 2B). After adjustment for age, BMI, and fibrinogen, 5‑fold higher odds of developing TE were seen in the patients with OAPS (OR, 5.26; 95% CI, 1.27–21.87). The Kaplan–Meier event‑free survival curves showed that reduced K_s (≤6.1 × 10^-9 cm²) was a predictor of TE in OAPS (Figure 3).