To explore the possible involvement of the proinflammatory and prothrombotic cytokine TNFalpha in APS by determining the plasma levels in patients and to test for association of TNFA promoter polymorphisms and HLA class II genotypes with both plasma TNFalpha and disease.We studied 83 Caucasoid patients with APS and two groups of healthy controls. TNFalpha levels were determined in plasma from 35 patients' and 21 controls using a highly sensitive sandwich ELISA. The full patient group was genotyped together with 95 ethnically matched healthy controls. -308 and -238 TNFA promoter polymorphisms were assessed by ARMS-PCR. HLA-DQB1, DQA1 and DRB1 genotypes were determined by PCR using sequence specific primers.TNFalpha levels were significantly higher in patients with APS than healthy controls (median 2.95 pg/ml [range 0.51-10.75] vs. 0.95 pg/ml [0.51-1.6], respectively; p = 0.0001). Frequencies of TNFA-308*2 genotype did not differ between patients and controls. In contrast, TNFA-238*A positive genotype was more frequent in APS patients with arterial thrombosis and pregnancy loss than in controls (OR 3.7 [95% CI 1.37-10.1], p = 0.007 and OR 3.95 [95% CI 1.3-11.7], p = 0.01; respectively). DQB1*0303-DRB1*0701 haplotype was associated with TNFA-238*A in the control group (OR 96.0 [95% CI 9.6-959], p0.0001) as well as in APS patient's group (OR 54.2 [95% CI 9.6-306.5], p0.0001).Raised plasma TNFalpha levels were found in patients with APS. As a prothrombotic and proinflammatory cytokine, TNFalpha may be involved in the development of clinical features of APS. The lack of correlation between the TNFA-238 polymorphism and plasma levels associated with disease suggests that the TNF genetic marker may only indirectly relate to protein levels by virtue of allelic association with a functional marker which may reside in the HLA class II region.