There is growing evidence for a role of factor XIII (FXIII) in vascular disease. FXIII measures were determined in (i) a nested case-control study from the Second Northwick Park Heart Study of 63 men with myocardial infarction (MI) and 124 age-matched controls and (ii) in a case-control study of 475 subjects with acute stroke and 461 controls followed up for 54 months for mortality. In both studies, measures of FXIII A- and B-subunit antigen, FXIII activity and prothrombin fragments (F1 + 2) were made. An in vitro model was used to investigate the effects of thrombin activity on FXIII A- and B-subunit antigen levels. In study 1, patients clinically free of coronary artery disease who later developed MI had lower adjusted FXIII A-subunit levels at recruitment (129.2%vs 113.3%, P = 0.007). In study 2, stroke patients with large vessel disease had lower A-subunit antigen levels (102.1%vs 127.2%, P < 0.001), but higher F1 + 2 levels (0.941%vs 0.753%, P < 0.05), than subjects with small vessel disease. Levels of FXIII A-subunit (100%vs 117%, P < 0.0001) were lower and F1 + 2 higher (1.020%vs 0.702%, P < 0.0001) in stroke patients who had died compared with those still alive at the end of the follow-up period. Low concentrations of FXIII A-subunit antigen predicted vascular outcome in otherwise healthy subjects and relate to both size of infarct and poor post-stroke survival in patients with acute ischaemic stroke. Low in vitro concentrations of FXIII A-subunit antigen wererelated to increased thrombin generation and, thus, increased risk of thrombotic events.