Thirty-three patients received at least one dose of fostamatinib during the study (Fig. 1); 24 completed period II, and 20 completed period III. Overall, 11/33 (33%) patients discontinued the study (adverse events [AEs] and lack of efficacy in 5 cases each, and withdrawal of consent in 1 case), which was lower than the global phase 3 trial (97/146 [66%]) [6]. There was no difference in baseline characteristics between patients who continued and discontinued fostamatinib throughout study period (Table S1).

Patient flow chart. A total of 34 patients were randomized to either the fostamatinib or the placebo group. Among them, 33 patients were treated with fostamatinib at least once. One patient who received the placebo withdrew from the study during period I. Twelve patients transitioned to a washout period. Twenty-two patients transitioned to period III. Two patients exited the study after period II, and 20 completed period III

The median (range) age of the 33 patients was 62 (25–81) years; 79% were female. The baseline platelet count was 19,000/µL (1,000–28,000/µL), and the ITP duration was 14 (1–41) years (Table S1). The duration of fostamatinib exposure was 771 (42–1184) days, and the median average daily dose was 238 (145–297) mg/day.

A platelet response > 50,000/µL (at two consecutive visits at least 28 days apart while receiving fostamatinib), which was predefined in this trial protocol [4], was achieved in 16/33 patients (48%) (Fig. 2). The median total duration of a platelet response > 50,000/µL was 589 (106–1,003) days. A platelet response > 30,000/µL was achieved in 18 patients (55%); the median total duration of such a response was 727 (57–1,170) days. No platelet overshoots (> 400,000/µL) were observed.

Swimmer plot with platelet response. Duration of platelet response > 50,000/µL: the maintenance start date was defined as the day after the start of study-drug administration, when the platelet count was 50,000/µL or higher for at least 28 consecutive days without the use of rescue medication. The maintenance end date was defined as the day when rescue medication was used or the day when the platelet count dropped below 50,000/µL for 28 or more consecutive days. The duration of the > 30,000/µL platelet response was defined in the same way. Data obtained after the first dose of fostamatinib are shown

Platelet counts > 50,000/µL were maintained at the end of the study in 13/16 patients. In two of the remaining three patients (#13 and #16), platelet counts were maintained > 30,000/µL without bleeding symptoms. The other patient (#10) experienced a drop in platelet counts to 10,000–20,000/µL during herpes zoster, but no bleeding was observed and the platelet count recovered spontaneously.

All 12 patients in the washout challenge had a decrease in platelet counts with the suspension of fostamatinib, but none experienced bleeding events (Fig. 3). Their platelet counts recovered during period III.

Platelet counts of patients who proceeded to the washout period and participated up to week 8 of period III. The dotted line indicates the transition to period III after the discontinuation of the washout period. Patient 2 withdrew from fostamatinib treatment at 14 days after the start of period III because of an elevated platelet count during the washout period

Patients were allowed to take concomitant ITP medications (glucocorticoids, azathioprine, or danazol) at a fixed dose during period I and a flexible dose after period I. Among 14 patients receiving glucocorticoid treatment at the beginning of period II, 3 underwent dosage reduction (prednisolone; 10–7.5 mg/day for 2 and 10–2 mg/day for 1), and 4 discontinued prednisolone (5, 5, 2.5, and 2 mg/day), owing to a treatment response to fostamatinib (Figure S1). Patients with a platelet count < 50,000/μL were allowed to use rescue medication. Rescue medication/treatment was administered to 13/33 (39%) patients, including glucocorticoids to 7, platelet transfusions to 5, and intravenous immunoglobulin to 4 patients. The major reasons for rescue medication/treatment were a lack of efficacy (six non-responders) and temporary use after the washout challenge (two patients).

All 33 patients experienced at least one AE (Table 1). The most common AEs (> 10% of cases) were diarrhea, hypertension, coronavirus disease 2019 (COVID-19), constipation, eczema, nasopharyngitis, and hepatic-enzyme elevation (Table S2). The majority of events were mildly (48%) or moderately (42%) severe. Treatment-related AEs that occurred in ≥ 10% of patients were diarrhea and hypertension (30% each). Most treatment-related AEs occurred before week 12 (Table S3).

No deaths occurred during the study. Serious AEs were observed in eight patients (24%), including COVID-19, cellulitis, gastroenteritis, herpes zoster, pericoronitis, thrombocytopenia, atrial fibrillation, diarrhea, large-intestinal polyps, autoimmune hepatitis, lumbar-spinal stenosis, and radial fracture (one each). Serious treatment-related AEs were observed in two patients (6%): thrombocytopenia and diarrhea (one each). AEs leading to drug withdrawal occurred in five patients (15%). Hepatic enzyme elevation (two patients) was the only AE leading to study-drug withdrawal of more than one patient.

The most frequent AEs were gastrointestinal disorders, hypertension, and hepatic enzyme elevation. Most of these events occurred within 12 weeks of treatment (Table 1 and Table S3). Only one of seven bleeding events, a mild case, was considered treatment-related. Thromboses, embolisms, and thromboembolisms were not observed during the study period. Infection was observed in 19 patients (58%), none of which were deemed related to fostamatinib. The infection was mild in 12 patients and moderate in 7 patients. None of the infections led to withdrawal from the study. COVID-19 occurred in 6 patients (18%). One patient was hospitalized for COVID-19 per their request, although the event was moderately severe. In that patient, COVID-19 occurred 592 days after the start of fostamatinib treatment; they were treated with remdesivir and recovered two days after admission. The investigator judged the disease unrelated with fostamatinib treatment because the patient had apparently had contact with a person with COVID-19 prior to the onset of the event.