Loss of antiphospholipid antibody positivity decreases the risk of recurrent thrombosis in thrombotic antiphospholipid syndrome.

OBJECTIVE: To establish the predictive value of antiphospholipid antibody (aPL) negativization on recurrent thrombosis in thrombotic antiphospholipid syndrome (APS). METHODS: Retrospective cohort study including all consecutive adult patients with APS followed at Hospital de Santa Maria, Lisbon, Portugal, up to December 2024. At diagnosis, all patients were positive for solid-phase aPL. Then, patients were categorized in two groups according to aPL during follow-up: (i) persistently aPL-positive; (ii) aPL-negativization defined as previous aPL-positive and current aPL-negative (two consecutive times, ≥12 months apart). Outcomes included recurrent thrombosis under antithrombotic therapy. Predictors of aPL-negativization and recurrent thrombosis were assessed using Cox regression analysis. RESULTS: Of the 116 included patients (female, 78.4%; primary APS, 72.1%; triple positivity, 40.5%), totalling 1084 patient years (PY), 31.9% became aPL-negative (3.4 events/100PY). The follow-up time was similar between aPL-positive and aPL-negativization groups (P = 0.681), and before and after aPL-negativization (P = 0.072). At diagnosis, triple aPL-positivity predicted against aPL-negativization, while primary APS predicted aPL-negativization. Forty-five patients (38.8%) experienced 68 recurrences (4.2 events/100PY). Age at APS onset, arterial thrombosis and persistent triple aPL-positivity predicted recurrence. Globally, aPL-negativization did not predict recurrence (HR 0.58, 95%CI 0.29-1.14). Among patients who became aPL-negative, the cumulative incidence of recurrent thrombosis did not differ before aPL-negativization (HR 0.52, 95%CI 0.26-1.06) but dropped by 88% after aPL-negativization (HR 0.12, 95%CI 0.05-0.58), even after adjusting for age, primary APS and type of initial thrombosis (P = 0.006). CONCLUSION: The risk of recurrent thrombosis drops significantly after aPL-negativization, highlighting the utility of serial aPL monitoring in clinical practice.