Clinical performance of obstetric antiphospholipid syndrome classification criteria: an ambispective cohort study

Abstract

Abstract Objectives To evaluate the clinical performance of the 2006 Sydney and 2023 ACR/EULAR classification criteria for obstetric antiphospholipid syndrome (OAPS), compare subsequent pregnancy outcomes across classifications and develop a supplementary classification method to further identify clinically relevant non-criteria OAPS (NC-OAPS) cases missed by the Sydney-derived NC-OAPS classification. Methods This single-center ambispective cohort included 1741 women of reproductive age enrolled from January 2020 to October 2024. Expert consensus was used to classify participants into OAPS, NC-OAPS and non-OAPS groups. The clinical performance of the 2006 Sydney criteria, 2023 ACR/EULAR criteria and Sydney-derived NC-OAPS classification were assessed. The subsequent pregnancy outcomes were followed up and compared across different groups. Moreover, a supplementary NC-OAPS classification method was developed and validated. Results The Sydney criteria identified expert-defined OAPS with sensitivity 0.806, specificity 0.991 and AUC 0.898. The 2023 ACR/EULAR criteria showed lower sensitivity (0.267) but near-perfect specificity (0.999). Live birth rates were similar among expert-defined OAPS, NC-OAPS and non-OAPS after treatment, but OAPS had higher rates of preterm birth, fetal growth restriction and severe pre-eclampsia/eclampsia. Patients fulfilling 2023 ACR/EULAR criteria had poorer pregnancy outcomes than those identified by expert consensus or the Sydney criteria. In the validation folds, the supplementary method yielded a sensitivity of 0.732 and a specificity of 0.921. Conclusion The 2006 Sydney criteria remain more suitable for clinical identification of OAPS, whereas the 2023 ACR/EULAR criteria may be useful for high-risk stratification. The supplementary classification method may improve recognition of NC-OAPS patients missed by existing criteria.