Prevalence, incidence proportion and independent predictors of thromboembolic events in adults with antiphospholipid syndrome: A systematic review with meta-analysis

Objectives: To synthesize data on the prevalence, incidence proportion, and independent predictors of thromboembolism in adults with definite antiphospholipid syndrome (APS) and those with ‘seronegative APS’. 

Methods: This systematic review and meta-analysis included relevant studies published between 1 January 2000 and 27 February 2022 retrieved through electronic database (MEDLINE, EMBASE, and Web of Sciences) and hand searches. Data was synthesized narratively and through random-effects aggregate meta-analyses. 

Results: We summarized data from 138 studies involving 21,963 adults with APS. The pooled point prevalence of ‘APS-classifying’ thromboembolic events in the definite APS population was 74.3% for ‘general thromboembolism’, 51.3% for venous thromboembolism (VTE), and 36.0% for arterial thromboembolism. The pooled period prevalence of catastrophic APS was 2.2% and that of recurrent events was 22.3% for ‘general thromboembolism’, 8.2% for arterial thromboembolism, and 14.3% for VTE. Pooled incidence proportions during definite APS follow-up were 20.4%, 15.2%, and 8.9% for ‘general thromboembolism’, arterial thromboembolism, and VTE, respectively. Based on very low to low level of evidence, lupus anticoagulant (LA) seropositivity and hypertension were major independent predictors of VTE and arterial thromboembolism, respectively. The pooled point prevalence of ‘APS-classifying’ thromboembolic events in the ‘seronegative APS’ population was 64.8% for ‘general thromboembolism’, 19.3% for arterial thromboembolism, and 31.9% for VTE. 

Conclusions: We found a high burden of thromboembolism among adults with Sapporo and Sydney criteria-based APS and those with ‘seronegative APS’. Based on very low to low level of evidence, LA is potentially an independent predictor of VTE and hypertension an independent predictor of arterial thromboembolism. Additional studies with standard methods are needed to improve the existing prediction models for thromboembolism in individuals with APS.