Non-criteria antiphospholipid antibody profiles and thrombotic outcomes in a cohort of patients with systemic lupus erythematosus Authors: Alistair Murray, Eric J Campbell, Ann Elaine Clarke, Megan R W Barber, Tania Pannu, Marvin J Fritzler, Michelle Jung, Yvan St Pierre, Leslie Skeith Published: 2024 ABSTRACT AbstractObjectivesAntiphospholipid syndrome (APS) is characterised by the presence of antiphospholipid antibodies (aPLs) and clinical outcomes of thrombosis and/or obstetric morbidity and is associated with systemic lupus erythematosus (SLE). IgG antiphosphatidylserine/prothrombin complex (aPS/PT), IgM aPS/PT and IgG anti-beta 2 glycoprotein 1-domain 1 (aβ2GP1-D1) are novel aPLs that have been associated with thrombosis; however, conclusive data are still lacking. It remains unclear how best to incorporate non-criteria autoantibodies into clinical decision-making. The aim of this study was to assess whether these novel aPLs were associated with an increased risk of thrombosis in patients with SLE.MethodsWe evaluated 341 patients enrolled in the SouThern Alberta Registry for Lupus EryThematosus database with SLE by the American College of Rheumatology or Systemic Lupus International Collaborating Clinics classification criteria. Medical records were reviewed between March 2006 and January 2021 for thrombotic events and serology results for lupus anticoagulant, IgG anticardiolipin, IgG anti-beta 2 glycoprotein 1 (aβ2GP1), IgG aPS/PT, IgM aPS/PT and IgG aβ2GP1-D1.ResultsAmong 341 patients with SLE, 59 (17%) met the revised Sapporo lab criteria, and of those 29 (49%) had a major thrombotic event (OR 3.5, 95% CI 1.9 to 6.3). Among 142 patients who had at least one positive non-criteria autoantibody, 45 (32%) had a major thrombotic event (OR 1.6, 95% CI 0.97 to 2.6). In a univariate analysis, the IgG aPS/PT and IgG aβ2GP1-D1 were associated with major and all thrombotic events. In a multivariate analysis that controlled for age, sex, prednisone use, SLE disease activity (Systemic Lupus Erythematosus Disease Activity Index—2K and the revised Sapporo lab criteria, among the non-criteria aPLs, only IgG aPS/PT was associated with an increased risk of a major thrombosis (OR 2.2, 95% CI 1.1 to 4.5).ConclusionsIn our multivariate analysis, IgG aPS/PT was associated with a modestly increased risk of thrombotic events.Keywords: Antiphospholipid Syndrome; Antibodies, Antiphospholipid; Lupus Erythematosus, Systemic WHAT IS ALREADY KNOWN ON THIS TOPIC Non-criteria antibodies against antiphosphatidylserine/prothrombin complex (aPS/PT) and anti-beta 2 glycoprotein 1-domain 1 (aβ2GP1-D1) have been associated with risk of thrombosis but there is conflicting evidence. Non-criteria antibodies against antiphosphatidylserine/prothrombin complex (aPS/PT) and anti-beta 2 glycoprotein 1-domain 1 (aβ2GP1-D1) have been associated with risk of thrombosis but there is conflicting evidence. WHAT THIS STUDY ADDS In a cohort of 341 patients with systemic lupus erythematosus (SLE), IgG aPS/PT was associated with thrombosis after adjusting for risk factors of age, sex, prednisone use, SLE disease activity and the revised Sapporo lab criteria. In a cohort of 341 patients with systemic lupus erythematosus (SLE), IgG aPS/PT was associated with thrombosis after adjusting for risk factors of age, sex, prednisone use, SLE disease activity and the revised Sapporo lab criteria. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY These data suggest IgG aPS/PT is associated with a modest increased risk of thrombosis. Further research is needed to confirm this finding and to better understand the interaction between IgG aPS/PT and the revised Sapporo criteria in the setting of thrombosis. These data suggest IgG aPS/PT is associated with a modest increased risk of thrombosis. Further research is needed to confirm this finding and to better understand the interaction between IgG aPS/PT and the revised Sapporo criteria in the setting of thrombosis. Introduction Antiphospholipid syndrome (APS) is a systemic autoimmune condition characterised by the presence of persistent autoantibodies directed against phospholipid-binding proteins and clinical outcomes of thrombosis and/or obstetric morbidity.1 There is an increased incidence of APS among patients with systemic lupus erythematosus (SLE).2 20–30% of patients with SLE will have positive antiphospholipid antibodies (aPLs), and in SLE cohorts with 8–10 years of follow-up, 15–20% of patients with positive aPLs experience a venous, arterial or small-vessel thrombotic event.3,6 The presence of aPLs is an independent predictor of mortality in patients with SLE.7 The 2006 revised Sapporo lab criteria, which are used in combination with clinical classification criteria to make a diagnosis of APS, include one or more of a positive lupus anticoagulant (LAC), IgG or IgM anti-beta 2 glycoprotein 1 (aβ2GP1), or IgG or IgM anticardiolipin (aCL). Antibodies must be persistently positive on repeat testing >12 weeks8. Based on the revised Sapporo criteria, antibodies against β2GP1 are considered positive at >99th percentile, while aCL is considered positive at >99th percentile or >40 GPU (immunoglobulin G phospholipid units).8 Some patients present with an APS-like phenotype but do not have a positive LAC, aβ2GP1 or aCL antibodies, a phenomenon known as seronegative APS (SNAPS).9 Novel aPLs are present in patients with SNAPS including IgG antiphosphatidylserine/prothrombin complex (aPS/PT), IgM aPS/PT and IgG anti-domain 1 β2GP1 (aβ2GP1-D1), but the significance of these non-criteria autoantibodies remains unclear.10 Updated classification criteria have recently been published by the American College of Rheumatology (ACR) and the European Alliance of Associations for Rheumatology (EULAR).11 12 Similar to the revised Sapporo criteria, the 2023 ACR/EULAR APS classification criteria include both clinical and laboratory domains. While new clinical entities have been added to the 2023 classification criteria to expand the definition of APS, the novel aPLs have not been added to the laboratory domain. There is interest in evaluating non-criteria autoantibodies, with the goal to better diagnose or prognosticate patients with APS or APS-like phenotypes. Prothrombin is a phospholipid binding protein that is detected in a complex with phosphatidylserine (PS/PT complex).13 Anti-PS/PT has been studied as a surrogate to the LAC, as a LAC can be an unreliable biomarker in severe illnesses where C-reactive protein can confound its detection, or in the presence of certain anticoagulants.14 15 In a systematic review of 10 studies including 1775 patients and 628 controls, aPS/PT autoantibodies were associated with arterial and venous thrombosis.16 Not all studies measured both IgG and IgM isotypes, and there were inconsistent findings among studies that did. A more recent systematic review reported on the prevalence of aPS/PT positivity among patients with APS and found a high rate of concomitant aPS/PT positivity (65%), especially among patients positive for LAC.17 aPLs are directed against phospholipid-binding proteins, of which, β2GP1 is one of the primary antigenic targets. The N-terminal domain I of β2GP1 is the primary binding site of pathogenic autoantibodies, which complex and interact with phospholipids and target cells including monocytes, endothelial cells and platelets.18 There is interest in studying the more specific IgG aβ2GP1-D1 biomarker, and IgG aβ2GP1-D1 has been shown to have a stronger association with risk of thrombosis than antibodies targeting other domains of β2GP1.19 There is less conclusive evidence for IgG aβ2GP1-D1 as an independent predictor of thrombosis, but it has been shown to aid in risk stratification among patients with established APS20 and among patients with LAC.19 Guo et al found that aβ2GP1-D1 positivity was associated with an increased risk of thrombotic events among patients with LAC.19 In contrast, other studies found that IgG aβ2GP1-D1 did not add diagnostic information over standard aβ2GP1 testing.21 Conclusive data quantifying the thrombotic risk for these non-criteria aPLs are still lacking, and it remains unclear how best to incorporate them into clinical decision-making. The aim of our study was to assess whether the non-criteria aPLs, IgG aPS/PT, IgM aPS/PT or IgG aβ2GP1-D1, were associated with an increased risk of a thrombotic event in a large cohort of patients with SLE after adjusting for factors including the established criteria aPLs. By understanding the risk of thrombosis in patients with non-criteria aPLs, we can provide further evidence if these antibodies should be incorporated in regular clinical practice or APS classification criteria. Methods We completed a retrospective cohort study of 341 patients with SLE who were enrolled in the SouThern Alberta Registry for Lupus EryThematosus (STARLET), a pre-existing prevalent observational cohort in Calgary, Alberta. Patients must be 18 years or older and fulfil the ACR 199722 or Systemic Lupus International Collaborating Clinics 201223 classification criteria for SLE, to be enrolled. Electronic medical records (EMRs) of patients enrolled in STARLET were reviewed from March 2006 (date of EMR start) to January 2021 (or study withdrawal or death if these occurred prior to 2021). Prednisone usage and Systemic Lupus Erythematosus Disease Activity Index (SLEDAI)—2K scores were available from the date of entry into the STARLET cohort. Available serology results for LAC, IgG aCL, IgG aβ2GP1, IgG anti-PS/PT, IgM anti-PS/PT and IgG aβ2GP1-D1 were recorded. Serology for IgM aCL and IgM aβ2GP1 is not routinely performed in Calgary, due to local established protocols that were based on prior institutional decisions according to available evidence. To confirm a LAC, two methods were required. The HemoSIL (Werfen, San Diego, California, USA) dilute Russell Viper Venom time screening assay, confirmation assay and mixing test were used (positive screen/confirm ratio >1.2). The second method used the SynthASil phospholipid reagent to screen for prolonged aPTT, and if prolonged, an actin FS aPTT and actin FSL aPPT were completed (positive FS/FSL ratio >1.5). Anti-domain 1 β2GP1 IgG was measured by chemiluminescence immunoassay (BioFlash, Werfen). All other aPLs were assessed using commercially available enzyme-linked immunosorbent assays (Werfen). The 99th percentiles were established as the cut-off for all aPL tests, including external validation by the kit manufacturer and internal validation through certification validation processes using both APS and control samples (Mitogen Diagnostics Corp, Calgary, Canada; Alberta Precision Laboratory, Calgary, Canada). The 99th percentile was <20 GPU for aCL and aβ2GP1, ≤20 for anti-domain 1 aβ2GP1 and ≤30 for IgG and IgM anti-PS/PT. Antibodies were not necessarily measured at the same time as the first-recorded thrombotic event, and the non-criteria aPLs were only tested between 2016 and 2020, but timing between aPL and events was recorded. EMRs were reviewed for the first recorded venous, arterial and small vessel thrombotic events. Using a province-wide EMR, all clinic visits, hospital records, lab values (including D-dimer and troponins) and imaging reports were reviewed from March 2006 to January 2021. Major thrombotic events were recorded and were based on the definition in the revised Sapporo criteria and included proximal and distal deep vein thrombosis (DVT), pulmonary embolism (PE) including subsegmental PE, and venous thrombosis of unusual sites including inferior vena cava, internal jugular vein and renal veins. Major arterial events included ischaemic stroke, myocardial infarction, peripheral arterial thromboembolism and left ventricular thrombi. Major venous and arterial thrombotic events were objectively confirmed by imaging records. Our expanded definition of any thrombotic event included major thrombotic events, as well as superficial vein thrombosis and transient ischaemic attack (TIA) diagnosed by a specialist. Cardiac valve abnormalities were not included. When there were discrepancies in the records relating to outcomes in data extraction (AM), there was a second review (LS) and discrepancies were resolved by consensus. In the case of simultaneous events (ie, DVT and PE), both events were recorded, and the participant was categorised as having experienced a major thrombotic event, irrespective of the number of events. ORs with CIs were calculated for risk of major thrombotic events and any thrombotic event by aPL profile; ORs were also calculated separately for these events by aPL profile among patients who met and who did not meet revised Sapporo lab criteria. Correlations between aPLs were examined using the Spearman correlation. A multivariate analysis examining the association between non-criteria aPLs and risk of major thrombotic events and any thrombotic event was conducted, controlling for revised Sapporo lab criteria, age, sex, mean prednisone use and mean SLEDAI 2K score. We computed adjusted mean values over the observed time period for both the SLEDAI 2K and prednisone use, and we included the last value observed for all patients as predictors. For prednisone use, the metric used was an average total dosage per month, in 100 mg of prednisone per month increments. In an exploratory analysis to understand the differential effect seen with IgG aPS/PT for patients who did or did not meet the revised lab Sapporo criteria, we performed multivariate regressions examining the association between non-criteria aPLs and risk of major and any thrombotic events, controlling for the above variables and including an interaction term for IgG aPS/PT and meeting the revised Sapporo criteria. We then performed multivariate analyses for each individual lab component of the revised Sapporo criteria with and without an interaction term for IgG aPS/PT and each of the aPLs. Data were analysed with Stata V.17 (StataCorp LLC, College Station, Texas, USA). Patients or members of the public were not involved in the design or conduct of this retrospective study. Results The available medical records of 341 patients with SLE were reviewed, with a median follow-up time of 13.4 years. Among the 341 patients, 314 (92.1%) were female with an average age of 46.7 years (SD ±16 years) (table 1). Of these, 161 (47%) had at least one major thrombotic event; venous events were the most common (97, 54% of all events), followed by arterial (58, 32% of all events), then small-vessel thrombosis (6, 3% of all events). All thrombotic events also included TIAs (16, 9% of all events) and superficial vein thrombosis (4, 2% of all events) (table 2). Table 1Demographic and clinical characteristicsDemographic characteristicsMean age (years, SD)46.7 (16)Female (n (%))314 (92.1%)Ethnicity (n (%)) White (Eastern Europe, Northern Europe, South Europe, West, Europe, North America, Australia/New Zealand, French Canadian, Acadia)205 (60.1%) Asian (East Asia, Southeast Asia, South Asia, Central Asia)78 (22.9%) Black (Africa, Caribbean, Hawaii, North America)17 (5.0%) Latin/Hispanic (Mexico, Central America, South America, Spain, USA)9 (2.6%) Arab/Middle Eastern (North Africa, Arabian Peninsula, Eastern Middle East, Northern Middle East)8 (2.3%) Indigenous (First Nations, Metis, Inuit)5 (1.5%) Pacific Islander (Fijian, Hawaii, Maori, Polynesia, Pacific Islands)2 (0.6%) Other12 (3.5%) Not reported5 (1.5%)Adjusted mean prednisone dose (mg, SD)0.93 (1.82)Adjusted mean SLEDAI-2K score (SD)3.46 (2.85)SLEDAI, Systemic Lupus Erythematosus Disease Activity Index Demographic and clinical characteristics SLEDAI, Systemic Lupus Erythematosus Disease Activity Index Table 2Thrombotic events among all patientsMajor thrombotic eventn% of Major events% of All eventsTotal161Venous thromboembolism976054 Deep vein thrombosis462925 Pulmonary embolism422623 Cerebral venous sinus thrombosis422 Other venous thrombosis*533Arterial thromboembolism583632 Ischaemic stroke311917 Myocardial infarction191210 Other arterial thrombosis†854Small vessel thrombosis643Other thrombotic eventn% of Other events% of All eventsTotal20 Transient Ischaemic attack16809 Superficial vein thrombosis4202*Other venous thrombosis includes: Rrenal vein thrombosis (3), inferior vena cava thrombosis (1), and internal jugular thrombosis (1).†Other arterial thrombosis includes: Pperipheral arterial embolic events (4), Ssplenic infarcts (3), and Lleft ventricular thrombosis (1). Thrombotic events among all patients Other venous thrombosis includes: Rrenal vein thrombosis (3), inferior vena cava thrombosis (1), and internal jugular thrombosis (1). Other arterial thrombosis includes: Pperipheral arterial embolic events (4), Ssplenic infarcts (3), and Lleft ventricular thrombosis (1). There were 341 patients who had revised Sapporo criteria antibodies measured. 59 (17%) met the revised Sapporo lab criteria with persistently positive LAC, aβ2GP1 or aCL on at least two occasions at least 12 weeks apart. Among the 59 patients who met revised Sapporo lab criteria, 29 (49%) had a major thrombotic event, and 33 (56%) had any thrombotic event (which included 4 additional TIA or superficial vein thrombotic events). Patients who met the revised Sapporo lab criteria were at increased risk of both major thrombotic events (OR 3.5, 95% CI 1.9 to 6.3) and any thrombotic events (OR 4.1, 95% CI 2.3 to 7.3) (table 3). There was an increased risk of both major thrombotic events and any thrombotic events with each of the individual antibodies of the revised Sapporo criteria, with details described in table 3. Table 3Association between aPL profiles and thrombotic eventsAntiphospholipid profilenMajor thrombotic eventAny thrombotic eventEvent(n, %)OR(95% CI)Event(n, %)OR(95% CI)Revised Sapporo lab criteria*Positive5929 (49) 3.5(1.9to 6.3) 33 (56) 4.1(2.3to 7.3) Negative28261 (22)67 (24)IgG aCLPositive4724 (51) 3.6(1.9to 6.8) 26 (55) 3.7(2.0to 6.9) Negative29466 (22)74 (25)IgG aβ2GP1Positive4122 (54) 4.0(2.0to 7.7) 24 (59) 4.2(2.1to 8.2) Negative30068 (23)76 (25)LACPositive3619 (53) 3.7(1.8to 7.5) 23 (64) 5.2(2.5to 10.8) Negative30571 (23)77 (25)Any non-criteria aPLPositive14245 (32)1.6 (0.97 to 2.6)52 (37) 1.8(1.1to 3.0) Negative17440 (23)42 (24)IgG aPS/PTPositive7032 (46) 3.2(1.8to 5.5) 34 (49) 3.0(1.7to 5.2) Negative24252 (21)59 (24)IgM aPS/PTPositive10832 (30)1.3 (0.8 to 2.1)38 (35)1.5 (0.9 to 2.5)Negative20452 (25)55 (27)IgG aβ2GP1-D1Positive2915 (52) 3.3(1.5to 7.2) 17 (59) 3.9(1.8to 8.4) Negative27867 (24)74 (27)Major thrombotic event: thrombotic event meeting Rrevised Sapporo criteria including venous, arterial, and small vessel thrombosis.Any thrombotic event includes superficial vein thrombosis and transient ischemicischaemic attacks, in addition to major thrombotic events.Bolded text denotes statistical significance based on the 95% confidence interval of the odds ratio.*Defined by one of: (1) LAC present in plasma, on two or more occasions at least >12 weeks apart, (2) aCL present in plasma in medium or high titertitre (>40 GPU or >99th percentile), on two or more occasions at least >12 weeks apart, (3) aβ2GP1 present in plasma >99th percentile, on two or more occasions at least >12 weeks apart.aCLanticardiolipinaPL, antiphospholipid antibody; aPS/PT, antiphosphatidylserine/prothrombin complex; aβ2GP1-D1, anti-beta 2 glycoprotein 1-domain 1GPUimmunoglobulin G phospholipid unitsLAClupus anticoagulant Association between aPL profiles and thrombotic events Major thrombotic event: thrombotic event meeting Rrevised Sapporo criteria including venous, arterial, and small vessel thrombosis. Any thrombotic event includes superficial vein thrombosis and transient ischemicischaemic attacks, in addition to major thrombotic events. Bolded text denotes statistical significance based on the 95% confidence interval of the odds ratio. Defined by one of: (1) LAC present in plasma, on two or more occasions at least >12 weeks apart, (2) aCL present in plasma in medium or high titertitre (>40 GPU or >99th percentile), on two or more occasions at least >12 weeks apart, (3) aβ2GP1 present in plasma >99th percentile, on two or more occasions at least >12 weeks apart. aCLanticardiolipinaPL, antiphospholipid antibody; aPS/PT, antiphosphatidylserine/prothrombin complex; aβ2GP1-D1, anti-beta 2 glycoprotein 1-domain 1GPUimmunoglobulin G phospholipid unitsLAClupus anticoagulant Non-criteria aPLs (IgG aPS/PT, IgM aPS/PT or IgG aβ2GP1-D1) were measured at least once in 316 (93%) of the patients, of which 142 (45%) were positive. Of all 341 patients, 246 (77%) had a non-criteria aPL measured at least twice, 52 (15%) had a non-criteria aPL measured once, and 25 (7%) did not have a non-criteria aPL measured. Among the 142 patients with positive non-criteria aPLs, 45 (32%) had a major thrombotic event, and 52 (37%) had any thrombotic event (which included 7 additional TIA or superficial vein thrombotic events). Positive non-criteria aPL was not associated with an increased risk of major thrombosis (OR 1.6, 95% CI 0.97 to 2.6) but was associated with an increased risk of any thrombotic event (OR 1.8, 95% CI 1.1 to 3.0) (table 3). All but one patient had positive aPLs and non-criteria aPLs measured within 5 years of a major thrombotic event. One patient had a remote history of DVT in the setting of SLE, but aPLs were only measured in 2014–2015 at the time of SLE autoantibody reassessment, and they were found to have a positive LAC and aβ2GP1 measured >12 weeks apart. A positive IgG aPS/PT was measured at least once in 70 patients and more than once (>12 weeks apart) in 41 patients. Among these 70 patients, 32 (46%) had a major thrombotic event and 34 (49%) had any thrombotic event. Patients with a positive IgG aPS/PT were at an increased risk of both a major thrombotic event (OR 3.2, 95% CI 1.8 to 5.5) and any thrombotic event (OR 3.0, 95% CI 1.7 to 5.2). A positive IgM aPS/PT was measured at least once in 108 patients and more than once (>12 weeks apart) in 58 patients. Among these 108 patients, 32 (30%) had a major thrombotic event, and 38 (35%) had any thrombotic event. Patients with a positive IgM aPS/PT were not at an increased risk of a major thrombotic event (OR 1.3, 95% CI 0.8 to 2.1), nor were they at an increased risk of any thrombotic event (OR 1.5, 95% CI 0.9 to 2.5). A positive IgG aβ2GP1-D1 was measured at least once in 29 patients and more than once (>12 weeks apart) in 20 patients. Among these 29 patients, 15 (52%) had a major thrombotic event, and 17 (59%) had any thrombotic event. Patients with a positive aβ2GP1-D1 were at an increased risk of both a major thrombotic event (OR 3.3, 95% CI 1.5 to 7.2) and of any thrombotic event (OR 3.9, 95% CI 1.8 to 8.4) (table 3). Among the 59 patients who met the revised Sapporo lab criteria, non-criteria aPLs were measured at least once in 55 (93%) patients, of which 35 (64%) had a positive IgG aPS/PT. Among these 35 patients, 24 (69%) had a major thrombotic event, and 25 (71%) had any thrombotic event. Patients who met revised Sapporo lab criteria and had a positive IgG aPS/PT were at an increased risk of both a major thrombotic event (OR 11.6, 95% CI 2.8 to 48.4) and any thrombotic event (OR 5.4, 95% CI 1.6 to 18.2). IgM aPS/PT and IgG aβ2GP1-D1 were not associated with increased risk of either major or any thrombotic event among patients who met the revised Sapporo lab criteria (online supplemental tables 1 and 2). Among the 282 patients who did not meet the revised Sapporo lab criteria, non-criteria aPLs were measured at least once in 261 (93%) patients, of which 94 (36%) were positive. There were 75 patients who had at least one positive non-criteria aPL and negative criteria aPL. There were 19 additional patients who had at least one positive non-criteria aPL and a single positive criteria aPL who did not meet the revised Sapporo lab criteria. Among the 75 patients with negative criteria aPL, 15 (20%) had a major thrombotic event, and 17 (22.7%) had any thrombotic event. None of the non-criteria aPLs were associated with either an increased risk of a major or any thrombotic event, although with small numbers of events and total patients reported (online supplemental table 2). There was a positive correlation between the revised Sapporo criteria aPLs and non-criteria aPLs (p<0.0001) (table 4). The IgG aPS/PT had a stronger correlation with LAC (Spearman=0.46), compared with the IgM aPS/PT (Spearman=0.37). The strongest correlation was between IgG aβ2GP1-D1 and IgG β2GP1 (Spearman=0.61) (table 4). Table 4Correlations between antiphospholipid antibodiesnSpearman correlationP valueIgG aPS/PT versus LAC3120.46p<0.0001IgM aPS/PT versus LAC3120.37p<0.0001aβ2GP1-D1 versus aβ2GP13040.61p<0.0001aPS/PT, antiphosphatidylserine/prothrombin complex; aβ2GP1-D1, anti-beta 2 glycoprotein 1-domain 1LAC, lupus anticoagulant Correlations between antiphospholipid antibodies aPS/PT, antiphosphatidylserine/prothrombin complex; aβ2GP1-D1, anti-beta 2 glycoprotein 1-domain 1LAC, lupus anticoagulant Among the 70 patients with a positive IgG aPS/PT, there were 43 patients who had at least one positive criteria aPL. Among those, 86.0% have at least one positive aCL; 76.7% at least one positive aβ2GPI, and 83.7% with at least one positive LAC. There were 81.4% who fulfilled the Sapporo lab criteria, with 72.1% having at least two positive aCL measured at least 12 weeks apart; 65.1% with at least two positive aβ2GPI measured at least 12 weeks apart, and 58.1% with two at least two positive LACs measured at least 12 weeks apart. In a multivariate analysis that controlled for the revised Sapporo lab criteria, age, sex, mean prednisone use (metric of prednisone 100 mg per month) and mean SLEDAI-2K, positive IgG aPS/PT independently increased the risk of a major thrombotic event (OR 2.2, 95% CI 1.1 to 4.5) but was not associated with a significant increased risk of any thrombotic event (OR 1.8, 95% CI 0.9 to 3.7). Positive IgM aPS/PT did not increase the risk of a major thrombotic event (OR 0.8, 95% CI 0.4 to 1.5) or any thrombotic event (OR 0.9, 95% CI 0.5 to 1.8). Positive IgG aβ2GP1-D1 was not associated with an increased risk of a major thrombotic event (OR 0.8, 95% CI 0.3 to 2.5) or any thrombotic event (OR 0.9, 95% CI 0.3 to 2.8) (table 5). Table 5Association of non-criteria antiphospholipid antibodies and thrombotic events, adjusting for the revised Sapporo criteria antiphospholipid antibodies, age, sex, mean prednisone use and mean SLEDAI-2K scoresCovariateRisk of major thrombotic eventOR (95% CI)Risk of any thrombotic eventOR (95% CI) Model1A:revised Sapporolabcriteria (no interaction term) IgG aPS/PT 2.2(1.1to 4.5) 1.8 (0.9 to 3.7)IgM aPS/PT0.8 (0.4 to 1.5)0.9 (0.5 to 1.8)IgG aβ2GP1-D10.8 (0.3 to 2.5)0.9 (0.3 to 2.8)Meets revised Sapporo lab criteria 4.1(1.7to 9.8) 4.9(2.1to 11.6) Age 1.03(1.01to 1.05) 1.04(1.02to 1.05) Female sex 0.3(0.1to 0.7) 0.3(0.1to 0.8) Prednisone use (by 100 mg/month) 1.2(1.0002to 1.3) 1.1 (0.96 to 1.3)SLEDAI-2K score1.0 (0.9 to 1.1)1.0 (0.9 to 1.2) Model 1B:revised Sapporolab criteriawith interaction term IgG aPS/PT0.9 (0.3 to 2.3)1.0 (0.4 to 2.4)IgM aPS/PT0.7 (0.4 to 1.5)0.9 (0.5 to 1.8)IgG aβ2GP1-D10.3 (0.1 to 1.3)0.5 (0.1 to 1.8)Meets revised Sapporo lab criteria1.0 (0.2 to 3.9)2.1 (0.7 to 6.6)Interaction termSapporo * IgG aPS/PT 31.7(4.4to 225.7) 8.9(1.6to 49.5) Age 1.04(1.02to 1.06) 1.04(1.02to 1.06) Female sex 0.3(0.1to 0.7) 0.3(0.1to 0.8) Prednisone use (by 100 mg/month) 1.2(1.04to 1.4) 1.1 (0.99 to 1.3)SLEDAI score1.0 (0.9 to 1.1)1.0 (0.9 to 1.1) Model2A: aCL (no interaction term) IgG aPS/PT 2.4(1.2to 4.9) 2.1(1.04to 4.2) IgM aPS/PT1.0 (0.5 to 1.8)1.2 (0.7 to 2.1)IgG aβ2GP1-D10.8 (0.3 to 2.7)1.0 (0.3 to 3.3)aCL 3.0(1.1to 8.0) 2.9(1.1to 7.5) Age 1.03(1.01to 1.05) 1.03(1.01to 1.05) Female sex 0.3(0.1to 0.7) 0.3(0.1to 0.9) Prednisone use (by 100 mg/month)1.1 (0.99 to 1.3)1.1 (0.95 to 1.3)SLEDAI score1.0 (0.9 to 1.1)1.0 (0.9 to 1.1) Model 2B: aCL with interaction term IgG aPS/PT1.5 (0.7 to 3.5)1.5 (0.7 to 3.3)IgM aPS/PT0.9 (0.5 to 1.7)1.2 (0.7 to 2.1)IgG aβ2GP1-D10.5 (0.1 to 1.9)0.7 (0.2 to 2.5)aCL0.8 (0.2 to 4.2)1.2 (0.3 to 5.0)Interaction term aCL * IgG aPS/PT 12.3(1.5to 100.8) 6.0 (0.9 to 39.7)Age 1.03(1.01to 1.05) 1.03(1.01to 1.05) Female sex 0.3(0.1to 0.7) 0.3(0.1to 0.9) Prednisone use (by 100 mg/month)1.2 (1.01 to 1.4)1.1 (0.97 to 1.3)SLEDAI score1.0 (0.9 to 1.1)1.0 (0.9 to 1.1) Model3A: aβ2GP1 (no interaction term) IgG aPS/PT 2.4(1.2to 4.9) 2.1(1.03to 4.2) IgM aPS/PT0.9 (0.5 to 1.7)1.1 (0.6 to 2.1)IgG aβ2GP1-D10.8 (0.2 to 2.7)1.0 (0.3 to 3.3)aβ2GP1 3.2(1.1to 9.3) 3.0(1.1to 8.6) Age 1.03(1.01to 1.05) 1.03(1.01to 1.05) Female sex 0.3(0.1to 0.7) 0.3(0.1to 0.9) Prednisone use (by 100 mg/month) 1.2(1.003to 1.3) 1.1 (0.97 to 1.3)SLEDAI score1.0 (0.9 to 1.1)1.0 (0.9 to 1.1) Model 3B: aβ2GP1 with interaction term IgG aPS/PT1.9 (0.9 to 4.2)1.8 (0.8 to 3.8)IgM aPS/PT0.9 (0.5 to 1.7)1.1 (0.6 to 2.1)IgG aβ2GP1-D10.6 (0.1 to 2.2)0.8 (0.2 to 2.9)aβ2GP11.3 (0.2 to 7.1)1.8 (0.4 to 8.1)Interaction term aβ2GP1 * IgG aPS/PT5.1 (0.6 to 41.5)2.7 (0.4 to 18.8)Age 1.03(1.01to 1.05) 1.03(1.01to 1.05) Female sex 0.3(0.1to 0.7) 0.3(0.1to 0.8) Prednisone use (by 100 mg/month) 1.2(1.01to 1.4) 1.1 (0.97 to 1.3)SLEDAI score1.0 (0.9 to 1.1)1.0 (0.9 to 1.1) Model4A: LAC (no interaction term) IgG aPS/PT 2.2(1.1to 4.6) 1.8 (0.9 to 3.6)IgM aPS/PT0.8 (0.4 to 1.6)0.9 (0.5 to 1.8)IgG aβ2GP1-D11.3 (0.4 to 3.6)1.4 (0.5 to 4.1)LAC 3.6(1.3to 9.7) 6.3(2.4to 17) Age 1.03(1.01to 1.05) 1.04(1.02to 1.06) Female sex 0.3(0.1to 0.8) 0.4(0.1to 0.9) Prednisone use (by 100 mg/month) 1.2(1.01to 1.4) 1.1 (0.98 to 1.3)SLEDAI score1.0 (0.9 to 1.1)1.0 (0.9 to 1.1) Model 4B: LAC with interaction term IgG aPS/PT1.9 (0.9 to 4)1.8 (0.8 to 3.8)IgM aPS/PT0.9 (0.5 to 1.7)0.9 (0.5 to 1.8)IgG aβ2GP1-D11.1 (0.4 to 3.2)1.4 (0.5 to 4.2)LAC1.0 (0.1 to 9.3) 6.9(1.5to 32.2) Interaction termLAC * IgG aPS/PT5.7 (0.5 to 67)0.9 (0.1 to 5.9)Age 1.03(1.01to 1.05) 1.04(1.02to 1.06) Female sex 0.3(0.1to 0.8) 0.4(0.1to 0.9) Prednisone use (by 100 mg/month) 1.2(1.01to 1.4) 1.1 (0.98 to 1.3)SLEDAI score1.0 (0.9 to 1.1)1.0 (0.9 to 1.1)Major thrombotic event: thrombotic event meeting revised Sapporo criteria including venous, arterial, and small vessel thrombosis.Any thrombotic event included superficial vein thrombosis and transient ischemicischaemic attacks, in addition to major thrombotic events.Note: All models include persistently positive criteria aPL measured at least >12 weeks apart aPS/PT, anti-phosphatidylserine/prothrombin complex; aβ2GP1-D1, anti-beta 2 glycoprotein 1-domain 1..Bolded text denotes statistical significance based on the 95% confidence interval of the odds ratio.aCLanticardiolipinaPS/PTantiphosphatidylserine/prothrombin complexaβ2GP1-D1anti-beta 2 glycoprotein 1-domain 1SLEDAI, Systemic Lupus Erythematosus Disease Activity Index—2K Association of non-criteria antiphospholipid antibodies and thrombotic events, adjusting for the revised Sapporo criteria antiphospholipid antibodies, age, sex, mean prednisone use and mean SLEDAI-2K scores Major thrombotic event: thrombotic event meeting revised Sapporo criteria including venous, arterial, and small vessel thrombosis. Any thrombotic event included superficial vein thrombosis and transient ischemicischaemic attacks, in addition to major thrombotic events. Note: All models include persistently positive criteria aPL measured at least >12 weeks apart aPS/PT, anti-phosphatidylserine/prothrombin complex; aβ2GP1-D1, anti-beta 2 glycoprotein 1-domain 1.. Bolded text denotes statistical significance based on the 95% confidence interval of the odds ratio. aCLanticardiolipinaPS/PTantiphosphatidylserine/prothrombin complexaβ2GP1-D1anti-beta 2 glycoprotein 1-domain 1SLEDAI, Systemic Lupus Erythematosus Disease Activity Index—2K While we observed that the IgG aPS/PT was associated with thrombosis after adjusting for other factors including meeting the revised Sapporo criteria, it was somewhat unexpected that in our stratified analysis, the IgG aPS/PT was not associated with thrombosis in those not meeting the revised Sapporo criteria (online supplemental table 2). Therefore, we completed exploratory regression analyses including interaction terms to better understand the independence of IgG aPS/PT, separate from the revised Sapporo criteria. The interaction between IgG aPS/PT and meeting the revised Sapporo criteria was significant for major thrombotic events (OR 31.7, 95% CI 4.4 to 225.7) and for any thrombotic events (OR 8.9, 95% CI 1.6 to 49.5), and neither IgG aPS/PT or the revised Sapporo criteria were independently associated with thrombosis when this interaction term was incorporated in the regression (table 5, Models 1A and 1B). We then explored this further by assessing each individual revised Sapporo aPL. The interaction between IgG aPS/PT and aCL was significant for major thrombotic events (OR 12.3, 95% CI 1.5 to 100.8) but not for any thrombotic event (OR 6.0, 95% CI 0.9 to 39.7). Neither IgG aPS/PT or aCL was associated with major thrombosis when this interaction term was incorporated into the regression, but when it was not included in the model for any thrombotic event (given it was not significant), IgG aPS/PT and aCL were associated with any thrombotic event (table 5, Models 2A and 2B). The interaction between IgG aPS/PT and aβ2GP1 was not significant (major thrombotic event: OR 5.1, 95% CI 0.6 to 41.5; any thrombotic event: OR 2.7, 95% CI 0.4 to 18.8) and when the interaction term was not incorporated into the regression, the IgG aPS/PT was associated with major and any thrombotic event (major thrombotic event: OR 2.4, 95% CI 1.2 to 4.9; any thrombotic event: OR 2.1, 95% CI 1.03 to 4.2) (table 5, Models 3A and 3B). Similarly, the interaction between IgG aPS/PT and LAC was not significant (major thrombotic event: OR 5.7, 95% CI 0.5 to 67.0; any thrombotic event: OR 0.9, 95% CI 0.1 to 5.9) and when the interaction was not included in the regression, IgG aPS/PT was associated with major thrombotic events (OR 2.2, 95% CI 1.1 to 4.6), and LAC was associated with major and any thrombotic events (table 5, Models 4A and 4B). Discussion LAC, aCL and aβ2GP1 antibodies are established risk factors for thrombosis and have been incorporated into classification criteria for APS.8 11 While IgG/IgM aPS/PT and IgG aβ2GP1-D1 were reported to be associated with thrombosis in some studies, it remains unclear if they add sufficient clinical value to be incorporated into APS classification criteria or inform clinical decision-making. In our cohort of 341 patients with SLE, we redemonstrated the frequency of the revised Sapporo criteria aPLs and their known increased risk of thrombosis. 17% of our cohort met revised Sapporo lab criteria as compared with 20% reported in the literature.24 In our cohort of patients meeting the revised Sapporo lab criteria, 49% had a major thrombotic event, which is higher than in other SLE cohort studies.3,6 In a modelling study of patients with SLE and aPLs, 50% are anticipated to have a thrombotic event over 20 years of follow-up.25 Patients with at least one positive IgG aPS/PT or IgG aβ2GP1-D1 were at an increased risk of major or any thrombotic events in the univariate analysis. In a multivariate analysis that controlled for additional factors including meeting the revised Sapporo lab criteria, among the non-criteria aPLs, only IgG aPS/PT was associated with an increased risk of a major thrombosis (OR 2.2, 95% CI 1.1 to 4.5). One strength of our cohort is the number who had non-criteria aPLs measured. While not the full sample size (93%), it has been our practice to measure these non-criteria aPLs at least once, leading to one of the largest cohorts with a rich dataset of both criteria autoantibodies, non-criteria autoantibodies and thrombotic outcomes. A systematic review by Sciascia et al16 found that the presence of aPS/PT was associated with an increased risk of arterial or venous thrombosis with an OR 5.11 (95% CI 4.2 to 6.3), although they did not separate IgG and IgM isotypes. This value is comparable to our finding that the presence of IgG aPS/PT increases the risk of major thrombosis with an OR 3.2 (95% CI 1.8 to 5.5), though IgM aPS/PT was not found to increase the risk of major thrombosis (OR 1.3, 95% CI 0.8 to 2.1). Similarly, a recent systematic review by Zhu et al found that the prevalence of IgG aPS/PT was associated with a history of thrombosis (OR values ranging from 2.7 to 11.0 between different studies).17 A 2012 retrospective cohort study of 158 patients with APS found that both IgG and IgM aPS/PT were associated with an increased risk of thrombosis (OR 4.1, 95% CI 1.4 to 11.7; OR 3.5, 95% CI 1.6 to 7.9, respectively).26 A 2017 multicentre retrospective study of 126 patients with APS found that the presence of IgG aPS/PT was associated with a history of thrombosis (OR 11.0, 95% CI 3.8 to 31.3).27 Antibodies binding β2GP1 via domain 1 are thought to be the most pathogenic aβ2GP1 antibodies.28 Recombinant domain 1 peptides have been proposed as potential therapeutic options for APS which have been shown to inhibit thrombosis in mouse models.28 29 While some studies have found that positive IgG aβ2GP1-D1 can predict an increased risk of thrombosis in patients who meet revised Sapporo criteria,17 30 others have found that IgG aβ2GP1-D1 does not add diagnostic power to the standard laboratory aPL tests.21 In our cohort, the presence of IgG aβ2GP1-D1 was associated with an increased risk of any thrombotic event among all patients in a univariate analysis (OR 3.9, 95% CI 1.8 to 8.4) but was not associated with an increased risk of thrombosis after controlling for other factors in our multivariate model including the revised Sapporo lab criteria. One challenge in interpreting a positive non-criteria aPL is when criteria aPL are also present, because of the correlation between the different antibodies. Anti-PS/PT has been proposed as a surrogate marker for LAC,31 although there are fewer data on IgG and IgM serotypes.16 In our cohort of patients with SLE, we observed correlations between IgG aPS/PT and LAC; between IgM aPS/PT and LAC; and between aβ2GP1-D1 and aβ2GP1. We identified that the correlation between IgG aPS/PT and LAC was stronger than the correlation between IgM aPS/PT and LAC. Vandevelde et al found that aPS/PT antibodies were independently associated with thrombosis when adjusting for aCL/aβ2GPI, but not when adjusted for LAC.32 To adjust for this possible confounding issue, we completed a multivariate analysis that controlled for age, sex, mean prednisone use, mean SLEDAI 2K scores and met the revised Sapporo lab criteria. The only non-criteria aPL associated with an increased risk of major thrombosis in the multivariate analysis was IgG aPS/PT, with an OR of 2.2 (95% CI 1.1 to 4.5). We examined the risk of thrombosis associated with individual aPLs among subgroups of patients who either did or did not meet the revised Sapporo lab criteria. Among patients who met the revised Sapporo lab criteria, IgG aPS/PT was the only non-criteria aPL associated with an increased risk of major (OR 11.6, 95% CI 2.8 to 48.4) and any (OR 5.4, 95% CI 1.6 to 18.2) thrombotic event. Among patients who did not meet the revised Sapporo lab criteria, none of the non-criteria aPLs was associated with an increased risk of thrombosis. Based on this discordant finding, we completed additional regression analyses including interaction terms, and found that while there was a significant interaction between IgG aPS/PT and meeting the revised Sapporo criteria, there was no significant interaction between IgG aPS/PT and the individual criteria aPLs, with the exception of aCL and major thrombotic events. Hence, when examining aPS/PT IgG and the criteria aPLs individually in models where the interaction term was not included (given it was not significant), aCL was associated with any thrombotic event, and IgG aPS/PT, aβ2GPI and LAC were associated with major and any thrombotic event. Given our limited sample size, further research is still needed to explore the different combinations of the individual revised Sapporo lab components with the positive and negative IgG aPS/PT, to make definitive statements of statistical independence. The 2023 ACR/EULAR APS classification criteria use a point-based system that combines data from clinical and laboratory domains. While the antibody types have not changed, there is now a point system of each antibody that has downgraded the IgM autoantibodies. In Calgary, IgM aCL and aβ2GPI are not done due to previously established protocols, so it is possible that the non-criteria aPL patients also had positive IgM aCL and/or IgM aβ2GPI present. IgM antibodies are only present in isolation <5% of the time33 34 and are not strongly associated with thrombotic outcomes.35,37 In a practice pattern survey of experts for the management of aPL-related stroke, ~25% of respondents omitted IgM autoantibodies from diagnostic testing, and ~50% would not treat with antithrombotic therapy in patients with isolated IgM autoantibodies.38 Research efforts are ongoing to further evaluate the role of isolated IgM autoantibody positivity and thrombotic risk. There were limitations to our study. Our chart review started in 2006 regardless of STARLET study enrolment, so it is possible that the thrombotic event occurred prior to the diagnosis of SLE and/or any aPL measured. The aPLs may not have been measured at the time of thrombosis. While we assessed for persistently positive revised Sapporo lab criteria antibodies, some patients had non-criteria antibodies assessed only once, and testing for these non-criteria antibodies was only performed at our institution between 2016 and 2020, therefore we did not capture changes in non-criteria aPLs, which are known to fluctuate overtime. For example, aβ2GP1-D1 specifically has been shown to fluctuate over time with levels being higher around the time of a thrombotic event and can be suppressed with the initiation of hydroxychloroquine.39 It is possible that the association of non-criteria aPL and thrombosis may have been stronger if they were required to be present on at least two occasions >12 weeks apart. We could not reliably determine the anticoagulation or antiplatelet status for our cohort of patients at the time of the thrombotic event, so we limited our analysis to the risk of first documented thrombosis and not the risk of recurrent thrombosis. It is possible that patients had a thrombotic event prior to 2006 that was not captured in our medical record, and that patients were already on aspirin for primary prophylaxis prior to their first documented thrombotic event. It is our general practice to recommend the use of low-dose aspirin for patients with SLE with revised Sapporo criteria antibodies in the absence of thrombosis, but generally not for the non-criteria antibodies, so this may have also impacted the risk of thrombosis between groups. Other confounding factors that this study did not control for are vascular risk factors, use of immunosuppression beyond prednisone and treatment with anticoagulation. There is a known association between vascular risk factors, such as hypertension, dyslipidaemia, diabetes, smoking and obesity, and risk of arterial thrombosis. Due to lack of data captured in our EMR, our study did not control for these comorbidities. However, we did adjust for selected SLE-related variables, that is, disease activity and prednisone dose, which are potentially associated with thrombosis. Lastly, we focus on thrombotic outcomes; obstetric APS was not assessed due to a lack of reliable documentation of obstetrical outcomes in our EMR. In summary, IgG aPS/PT and IgG aβ2GP1-D1 were associated with a risk of thrombotic events in a large retrospective cohort of patients with SLE. However, both non-criteria aPLs were found to be correlated with established aPLs. In our multivariate analysis, only the IgG aPS/PT non-criteria autoantibody was associated with a modestly increased risk of any thrombotic events, when controlling for sex, age, mean prednisone use, mean SLEDAI-2K score and the revised Sapporo criteria aPLs. Additional studies are needed to confirm these findings and to better understand the interaction between IgG aPS/PT and the revised Sapporo lab criteria in the setting of thrombosis. supplementary material Acknowledgements We would like to thank Haiyan Hou and Chynace Lambalgen (University of Calgary) for technical assistance, and all participants included in the STARLET database. Footnotes Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Not applicable. Ethics approval: This study was conducted in accordance with the Canadian Tri-Council Policy Statement for the ethical conduct of research involving humans and was approved by the University of Calgary Research Ethics Board (REB#13-0425). Participants gave informed consent to participate in the study before taking part. Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research. Data availability free text: Please contact the authors for data requests. Contributor Information Alistair Murray, Email: alistair.murray93@gmail.com. Eric J Campbell, Email: ejcampbe@ucalgary.ca. Ann Elaine Clarke, Email: aeclarke@ucalgary.ca. Megan R W Barber, Email: mrwbarbe@ucalgary.ca. Tania Pannu, Email: tania.pannu@albertahealthservices.ca. Marvin J Fritzler, Email: marv.fritzler@mitogendx.com. Michelle Jung, Email: michelle.jung@albertahealthservices.ca. Yvan St. Pierre, Email: yvan.stpierre@rimuhc.ca. Leslie Skeith, Email: laskeith@ucalgary.ca. Data availability statement Data are available upon reasonable request. References Associated Data Supplementary Materials Data Availability Statement Data are available upon reasonable request.