The main criticism of De Stefano and colleagues pertains to our suggestion to use bone marrow biopsy in the diagnostic pathway of chronic myeloproliferative disorders (CMPDs) associated with splanchnic vein thrombosis (SVT), together with the search of the JAK2 V617F mutation and clonality assays. Our recommendation is not new, and reinforces a well-documented literature.1, 2 Bone marrow biopsy has for a long time been part of the diagnostic routine in our and other institutions for these patients, as documented by the fact that in our retrospective series, a bone marrow biopsy was performed in 74 of 93 cases.3 In our consideration for including bone marrow biopsy in the diagnostic workup of these patients, risk-benefit was not taken into account, because it is considered a trivial argument. The rate of adverse events for a bone marrow biopsy cited by De Stefano and colleagues, derived from 19,259 procedures (both bone marrow biopsies and aspirates) performed from 2001 to 2003 in the United Kingdom, is 0.08%.4 However, because only 3 of these patients had serious events, we should expect that 1.3 patients in 10,000 would risk serious events from a bone marrow biopsy. This is the reason why hematologists consider bone marrow biopsy a very safe procedure. Our standard policy before performing bone marrow biopsy in patients on oral anticoagulant therapy is to stop this therapy 3 days before and to start low weight molecular heparin (LWMH) until the evening before the procedure, to skip its administration in the morning, and to start again LWMH in the evening together with coumarin derivatives, continuing until the target International Normalized Ratio is reached. This policy is safe, and we observed no adverse events.
Pertaining to the criticism to our suggestion that the presence of the JAK2 mutation or a bone marrow biopsy suggestive for a CMPD should be considered per se a prothrombotic state, several studies indicate that CMPD patients carrying the JAK2 mutation have a higher risk of thrombotic complications5, 6 and such risk is probably due to increased blood cell production and activation, irrespective of the number in the peripheral blood. This is particularly true in patients with extrahepatic portal vein obstruction or Budd-Chiari syndrome in whom typical peripheral blood changes may be lacking, in part or totally, because of increased plasma volume or hypersplenism caused by portal hypertension. The vast majority of CMPD patients in our study, either JAK2-positive or not, as well as in a similar study,7 indeed lacked the conventional hematological criteria for suspicion of myeloproliferation. These arguments and the experience of a very high risk of re-thrombosis, either splanchnic or not, argue in favor of life-long anticoagulation in patients with SVT and CMPD. Pertaining to the possible indication for cytoreduction, we are aware that there is no evidence to recommend cytoreduction instead of (or together with) anticoagulants in patients with CMPDs with SVT who lack the conventional hematological criteria. It appears however that the conventional criteria for starting cytoreductive treatment in these atypical patients are inconsistent and should be reset at a much lower cut-off of blood peripheral counts, because the rationale should be the suppression of the abnormal clone responsible for myeloproliferation. This approach is perhaps particularly indicated in JAK2-positive patients, whose propensity for thrombosis appears to be higher than that of JAK2-negative patients.