We read with interest the recent paper by Peverill et al. [1] concerning the effects of percutaneous balloon mitral valvuloplasty (PBMV) on levels of prothrombin fragment 1j2 (F1j2) in mitral stenosis. They report that, overall, venous F1j2 levels increased, whereas atrial levels did not change after PBMV. With subgroup analyses, however, venous levels of F1j2 were significantly increased and left atrial levels decreased in the subgroup of 15 patients with increased regional left atrial thrombin generation, whereas both venous and left atrial F1j2 levels increased in the 22 patients with normal left atrial thrombin generation [1].They postulated that the reduction in left atrial F1j2 levels in the former group could either be due to increased washout of preformed left atrial F1j2 into the systemic circulation or a decrease in left atrial blood stasis [1]. If this hypothesis is correct, similar changes (i.e. a reduction in left atrial F1j2 levels) would also have been observed in the cohort of patients with normal left atrial thrombin generation, but this was not apparent.An alternative explanation for the reduction in left atrial F1j2 levels could be related to some mitral regurgitation after the PBMV procedure, but their data suggest that this appears to have been unlikely [1]. Indeed, mitral regurgitation is known to be associated with reduced indices of hypercoagulability [2]. The observations by Peverill et al. [1] are also limited by samples taken immediately pre-PBMV and immediately afterwards. It would be interesting to know whether the apparent left atrial reduction in F1j2 levels (albeit in the small subgroup with increased regional left atrial thrombin generation) is a sustained phenomenon or not. Furthermore, 19 patients were taking warfarin preprocedure and 18 patients were not, and whereas all patients had their warfarin stopped and international normalized ratio levels were ' in the normal range ', the possibility remains that warfarin may have affected levels of some of their measured indices. Finally, 13 of their 37 patients had atrial fibrillation, and this common arrhyth-