Patients randomized to the BAT arm could crossover to ruxolitinib treatment at any time during the study upon a protocol-defined progression event. Patients in the ruxolitinib arm who had a protocol-defined progression event could continue receiving ruxolitinib in the extension phase at any time during the study if, in the investigator’s opinion, they were still receiving a benefit from ruxolitinib treatment. Progression events that qualified for the crossover and extension phases included the need for splenectomy and progressive splenomegaly as defined by a 25% increase in spleen volume compared with the on-study nadir (including baseline). After the primary analysis, the study protocol was amended (amendment 5) in January 2011 to allow all patients to enter the extension phase, including those who did not meet the criteria for progression.

Six patients crossed over from BAT to ruxolitinib prior to protocol amendment 5 without experiencing qualifying progression events (5 patients discontinued due to the protocol deviation and 1 patient discontinued due to other reason).

Other reasons for discontinuation in the ruxolitinib arm included patients who underwent stem cell transplant (n = 5), interruption of study medication for >8 weeks (n = 2), lack of efficacy (n = 2), meeting protocol-defined imaging discontinuation criteria (n = 2), investigator decision (n = 1), diagnosis of lung cancer with the start of chemotherapy treatment (n = 1), unspecified safety event (n = 1), and modest spleen response (n = 1). Other reasons in the BAT arm included stem cell transplant (n = 2), investigator decision (n = 2), patient decision (n = 2), splenic irradiation (n = 1), hematemesis and thrombocytopenia (n = 1), unwillingness to undergo magnetic resonance imaging (n = 1), enrollment in a ruxolitinib compassionate use program (n = 1), splenectomy (n = 1), initiating treatment with hydroxyurea (n = 1), and thrombocytopenia as sign of disease progression (n = 1).