CADTH Reimbursement Review

Fostamatinib (Tavalisse)

Sponsor: Medison Pharma Canada Inc.

Therapeutic area: Chronic immune thrombocytopenia

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Stakeholder Input

Clinical Review

Executive Summary

An overview of the submission details for the drug under review is provided in Table 1.

Introduction

Immune thrombocytopenia (ITP) is an autoimmune disease characterized by low platelet counts and increased bleeding risk.¹ Primary ITP is not triggered by a specific condition or event, while secondary ITP is caused by or associated with another condition. Chronic ITP lasts 12 months or more after the initial diagnosis while persistent ITP lasts 3 to 12 months after initial diagnosis. A 2010 narrative review of international studies suggested that the incidence of ITP among adults is approximately 3.3 per 100,000 per year while the prevalence is 10 per 100,000. Patients with ITP may experience bleeding (mild, severe, or critical) and fatigue, and have a reduced health-related quality of life (HRQoL) due to the disease and its treatment. Treatment is generally indicated when the platelet count is low (e.g., < 20,000/µL or 30,000/µL) and/or if the patient is experiencing bleeding.¹ The main goals of therapy in ITP are to prevent severe or critical bleeding, reduce or eliminate patients’ symptoms, minimize adverse effects from treatments, and ultimately improve patient HRQoL.² Recommended treatments that target platelet levels above 20,000/µL to 30,000/µL appear to reduce the risk of major bleeding.^2,3 Corticosteroids or IV immunoglobulin (IVIG) are recommended by guidelines as first-line therapy.^2,3 Patients may not respond to these treatments or relapse. In such patients, several subsequent-line treatments are available, such as a splenectomy; rituximab; thrombopoietin receptor agonists (TPO-RAs) such as romiplostim and eltrombopag; fostamatinib; or immunosuppressants.^2,3 The optimal sequence of second- and subsequent-line therapies is often unclear due to a lack of comparative efficacy and safety data, patient heterogeneity, and access and/or reimbursement issues.^2,3 Not all patients respond to treatment with second- or third-line treatment options. Patients can also become refractory to treatment options or relapse after achieving remission.^2,3 Chronic ITP is therefore characterized by a chronic relapsing course and multiple lines of therapy over time.

Fostamatinib is indicated for the treatment of chronic ITP in adult patients who have had an insufficient response to other treatments.⁴ It reduces destruction of platelets via inhibition of spleen tyrosine kinase.⁴ Fostamatinib is initiated at a dosage of 100 mg twice daily and is taken orally.⁴ If the platelet count has not increased to at least 50,000/µL after 4 weeks, then the dosage can be increased to 150 mg twice daily.⁴ Fostamatinib underwent an expedited review at Health Canada.

The objective of this report was to perform a systematic review of the beneficial and harmful effects of fostamatinib for the treatment of thrombocytopenia in adult patients with chronic ITP who have had an insufficient response to other treatments.

Stakeholder Perspectives

This section summarizes input provided by the patient groups who responded to CADTH’s call for patient input and from clinical experts consulted by CADTH for the purpose of this review.

Patient Input

One patient group submission, prepared by the Platelet Disorder Support Association (PDSA), was received for this review. How the data were collected to inform the submission was not described; however, patient experiences with fostamatinib were gathered from the PDSA’s Facebook group. The patient input submission suggested that patients with ITP are fearful of life-threatening bleeding, face physical and emotional consequences from their disease (e.g., fatigue, anxiety, and depression), and restrict their activities because of their disease. The submission also suggested that ITP and its treatment interfere with daily life and negatively affect HRQoL. Patients are often more concerned with managing symptoms and improving HRQoL than with platelet counts. Myriad treatment options are available to manage ITP, and it is difficult to predict who will respond to a particular treatment and who will develop resistance to a treatment over time. In addition, patients may not be able to afford or access available options. It is therefore important that patients have options available in case they do not respond to a therapy, the therapy stops working, or they experience bleeding.

Clinician Input

Input From Clinical Experts Consulted by CADTH

The clinical experts consulted by CADTH suggested that standard first-line therapy for ITP includes corticosteroids, and IVIG is often added when an immediate increase in platelets is required, although the effect of IVIG is often transient. The experts noted that a significant proportion of patients will not respond to steroids and, of those that do, many will relapse once steroids are tapered. At this point, a splenectomy is the traditional second-line therapy if the patient is a suitable candidate. More recently, rituximab has emerged as an alternative second-line therapy. If both a splenectomy and rituximab have failed (or are contraindicated), a large number of third-line therapies are available, including immunosuppressant medications such as azathioprine or cyclophosphamide, or TPO-RAs such as eltrombopag or romiplostim. There is little evidence to guide the selection of second- or third-line therapy, and decisions depend on both local reimbursement considerations and patient-specific factors.

The clinical experts consulted by CADTH stated that treatment goals are to reduce bleeding and prolong life. Increasing the platelet count is generally considered to be a reasonable surrogate for both goals. Improving HRQoL is also an important goal.

The clinical experts emphasized that not all patients respond to available therapies, and even if remission is achieved, long-term remission is not guaranteed. The clinical experts noted how accessibility to appropriate second- and third-line therapies can be a challenge, as not all options are reimbursed in each province or because reimbursement criteria differ across provinces. Administration of existing therapies can also be a challenge and there are adverse effects with existing treatments. Therapies with demonstrated efficacy, convenience of administration, and a low risk of adverse effects would therefore fill an unmet need for treatment of ITP.

The clinical experts stated that contemporary ITP guidelines suggest that, in general, a splenectomy or rituximab can be considered as second-line therapy. Several third-line options are available; however, the comparative efficacy of these agents is unclear. It can therefore be difficult to determine the best treatment option for a particular patient, and there is often no single clearly defined treatment pathway. Decisions end up being driven largely by access. It is challenging to identify the optimal place for fostamatinib in a therapeutic algorithm. The clinical experts noted that rituximab or a splenectomy are reasonable second-line choices (TPO-RAs may also be considered second-line choices in some patients). The safety profile of fostamatinib and the fact that it is administered orally suggest it may be considered a reasonable third-line therapy rather than reserved for patients who have failed or do not have access to TPO-RAs, as has been proposed by the sponsor. However, regardless of where it sits in the therapeutic algorithm, it would be advantageous for clinicians if fostamatinib was available as a treatment option for specific patients. The clinical experts noted that the ITP population is heterogenous, and the available data and current understanding of ITP pathophysiology make it impossible to determine who will respond best to fostamatinib and who are most susceptible to adverse effects.

Bleeding is an important outcome in the treatment of ITP, and ultimately any treatment should reduce the occurrence of clinically important bleeding while improving HRQoL. In practice, clinicians rely on platelet response, which is assumed to reduce the risk of clinically relevant bleeding and, as a secondary benefit, reduce the need for rescue therapy. In general, an increase in platelet count can be seen as early as 2 weeks into treatment with fostamatinib, although some patients may not respond until week 12. If a response is observed, clinicians would likely continue to use the treatment long-term with monthly monitoring. A sustained response would generally be considered a platelet count of 30,000/µL to 50,000/µL for the duration of a treatment cycle (e.g., 24 weeks). If a response has not been seen by approximately 24 weeks, most clinicians would likely consider the treatment to have failed and would discontinue it. If there are issues related to safety or tolerability, treatment would generally be discontinued earlier, particularly if it is affecting a patient’s HRQoL.

Clinician Group Input

A group of 19 Canadian hematologists submitted input on fostamatinib. The clinician group submission echoed the opinions of the expert panel. The clinician group submission suggested that fostamatinib would be likely used after first-line therapy as second-line or subsequent-line therapy. The clinician group submission reported that fostamatinib would be used as a single drug after first-line therapy has failed. Fostamatinib was described as an alternative to other second- and subsequent-line therapies and should be considered before a splenectomy, immunosuppressive drugs, and rituximab and its biosimilars, and should be used at a line of therapy similar to that of maintenance treatments such as TPO-RAs. The clinician group submission stated that patients earlier in their ITP disease course may respond better to fostamatinib. Using it as a second-line treatment may therefore offer advantages, such as limiting exposure to complications or toxicities from other drugs. However, the greatest need is still in patients who have relapsed multiple times despite treatment.

Drug Program Input

Drug programs asked whether concomitant therapy (e.g., corticosteroids or danazol) is required with fostamatinib as some patients from relevant clinical trials used concomitant therapy. The clinical experts noted that concomitant therapy is likely not necessary to continue and could often be discontinued if and when it is considered unnecessary. Drug programs also asked how the time to response for fostamatinib compares to other available treatments. The clinical experts reported that fostamatinib has a similar time to response as TPO-RAs and is faster than rituximab. The drug programs inquired about the threshold for initiating treatment in cases of ITP and the clinical experts clarified that the usual threshold is less than 30,000/µL. The drug programs also asked if fostamatinib should only be reimbursed for people with ITP for more than 3 years based on subgroup analyses in relevant clinical trials. The clinical experts noted that trials were likely underpowered to detect a response in subgroups and indicated that it would not be appropriate to exclude patients from reimbursement based on a duration of ITP of less than 3 years. Drug programs asked if public payers should only fund fostamatinib after the patient has failed all second-line therapies (splenectomy, rituximab, and TPO-RAs). The clinical experts suggested that, given the lack of comparative evidence among treatment options, it is challenging to put fostamatinib in front of established second-line therapies such as rituximab or a splenectomy (unless there are specific contraindications), and that it may be more reasonable to position fostamatinib at a level of treatment similar to that of TPO-RAs. The drug program asked what a typical treatment course of fostamatinib would be and the clinical experts suggested that, because it has not been used extensively in practice to date, there is no recognized typical course of treatment. The drug programs also asked if, based on adverse event (AE) data from the relevant clinical trials, fostamatinib can be considered to be poorly tolerated. The clinical experts stated that, because the placebo group in the trials also experienced high rates of AEs, they cannot say that fostamatinib is poorly tolerated.

Clinical Evidence

Pivotal Studies and Protocol-Selected Studies

Description of Studies

Two identically designed 24-week double-blind randomized controlled trials (RCTs), FIT1 (N = 76)⁵ and FIT2 (N = 74),⁶ evaluated the efficacy and safety of fostamatinib versus placebo in patients with primary ITP for more than 3 months who had received at least 1 previous ITP treatment and had a baseline platelet count below 30,000/µL in at least 3 counts in the preceding 3 months. The FIT1 trial was conducted in Australia, Canada, 4 countries in Europe (Denmark, Hungary, Italy, the Netherlands), the UK, and the US, while the FIT2 trial was conducted in 8 countries in Europe (Austria, Bulgaria, Czech Republic, Germany, Norway, Poland, Romania, Spain). In the FIT1 trial, 51 patients were randomized to fostamatinib and 25 to placebo, while in the FIT2 trial, 50 patients were randomized to fostamatinib and 24 to placebo. The primary efficacy end point in both trials was achievement of stable platelet response, defined as a platelet count of least 50,000/µL at 4 of the last 6 study visits between weeks 14 and 24. These trials also measured the use of rescue therapy, bleeding-related serious adverse events (SAEs), and HRQoL (via the Short Form [36] Health Survey [SF-36]), along with harms.

In the FIT1 trial, the mean age was 57 years (standard deviation [SD] = 18) in the fostamatinib group and 53 years (SD = 16) in the placebo group. In the fostamatinib group, 59% of patients were female compared to 68% in the placebo group. The mean duration of ITP was 13 years (SD = 14) in the fostamatinib group versus 9 years (SD = 10) in the placebo group. Patients in the fostamatinib group had used a median of 5 prior ITP treatments (range = 1 to 10) while patients in the placebo group had used a median of 3 (range = 1 to 9). More patients in the placebo group had used steroids (100%) and TPO-RAs (60%) compared to the fostamatinib group (90% for steroids and 51% for TPO-RAs). In the FIT1 and FIT2 trials, the rate of concomitant steroid use was higher in the placebo group (56% in the FIT1 trial and 63% in the FIT2 trial) compared to the fostamatinib group (37% in the FIT1 trial and 44% in the FIT2 trial). In the FIT1 trial, the rates of prior splenectomy were similar (40%) in both groups. In the FIT2 trial, the mean age was 49 years (SD = 15) in the fostamatinib group and 50 years (SD = 17) in the placebo group. In the fostamatinib group, 62% of the patients were female compared to 54% in the placebo group. The mean duration of ITP was 12 years (SD = 13) in the fostamatinib group versus 11 years (SD = 8) in the placebo group. Patients in both groups had used a median of 3 previous ITP treatments (range = 1 to 10). The rate of previous individual ITP medication use was similar between groups. The rates of a prior splenectomy was higher in the placebo group (38% versus 28% in the fostamatinib group).

Efficacy Results

In the FIT1 trial, 18% of patients in the fostamatinib group experienced a stable platelet response compared to 0% in the placebo group (risk difference [RD] = 18%; 95% confidence interval [CI], 7.2 to 28; P = 0.026). In the FIT2 trial, 18% of patients in the fostamatinib group experienced a stable platelet response compared to 4% in the placebo group (RD = 14%; 95% CI, 0.5 to 27; P = 0.15), a difference that was not statistically significant.

In the FIT1 trial, 31% of patients in the fostamatinib group required rescue therapy before week 10 compared to 44% of patients in the placebo group. After week 10, rescue therapy was required for 14% of patients in the fostamatinib group compared to 28% of the placebo group. In the FIT2 trial, 18% of patients in the fostamatinib group required rescue therapy before week 10 compared to 29% of patients in the placebo group. After week 10, only 2% of patients in the fostamatinib group required rescue therapy compared to 21% in the placebo group. In the FIT1 trial, ||||| of patients in the fostamatinib group experienced a bleeding-related SAE compared to |||||| in the placebo group. In the FIT2 trial, ||||| of patients in the fostamatinib group experienced a bleeding-related SAE compared to ||||| in the placebo group. No statistical testing was applied to these outcomes.

For the quality-of-life outcome, no differences in SF-36 scores were evident between the fostamatinib and placebo groups at any time point in the FIT1 trial. At week 24, there was |||||||||| providing SF-36 data in the placebo group and ||||| patients in the fostamatinib group. In the FIT2 trials, no differences in SF-36 scores were evident between the fostamatinib and placebo groups at week 12 or week 24. At week 24 in the FIT2 trial, there were ||||||||||| providing SF-36 data in the placebo group and ||||| patients in the fostamatinib group. The effect of fostamatinib on HRQoL is unclear from the FIT1 and the FIT2 trials.

Both the FIT1 and FIT2 trials conducted subgroup analyses for the primary efficacy end point. In the FIT1 trial, among patients with prior TPO-RA treatment, 15% of the fostamatinib group experienced a stable platelet response compared to 0% of the placebo group (RD = 15%; 95% CI, 1.5 to 29). Among patients without prior TPO-RA treatment, 20% of the fostamatinib group experienced a stable platelet response compared to 0% of the placebo group (RD = 20%; 95% CI, 4.3 to 36). In the FIT2 trial, among patients with prior TPO-RA treatment, 15% of the fostamatinib group experienced a stable platelet response compared to 0% of the placebo group (RD = 15%; 95% CI, −0.6 to 31). Among patients without prior TPO-RA treatment, 20% of the fostamatinib group experienced a stable platelet response compared to 7% of the placebo group (RD = 13%; 95% CI, −6.8 to 33). In the FIT1 trial, among patients with a prior splenectomy, 15% of the fostamatinib group experienced a stable platelet response compared to 0% of the placebo group (RD = 15%; 95% CI, −0.6 to 31). Among patients without a prior splenectomy, 19% of the fostamatinib group experienced a stable platelet response compared to 0% of the placebo group (RD = 19%; 95% CI, 5.4 to 33). In the FIT2 trial, among patients with prior splenectomy, 21% of the fostamatinib group experienced a stable platelet response compared to 0% of the placebo group (RD = 21%; 95% CI, −0.1 to 43). Among patients without prior splenectomy, 17% of the fostamatinib group experienced stable platelet response compared to 7% of the placebo group (RD = 10%; 95% CI, −7.5 to 28).

The following outcomes identified in the protocol were not reported in either the FIT1 or FIT2 trial: duration of response, symptoms, hospitalizations, or emergency room visits.

Harms Results

In the FIT1 trial, among patients in the fostamatinib group, the most common AEs (≥ 5%) were diarrhea (41%), nausea (29%), increased alanine transaminase (ALT) (18%), increased aspartate transaminase (AST) (16%), headache (14%), dizziness (18%), epistaxis (18%), fatigue (12%), and hypertension (26%). The most common AEs in the placebo group were diarrhea (16%), headache (24%), dizziness (16%), epistaxis (16%), and dyspnea (12%). In the FIT2 trial, the most common AEs in the fostamatinib group were diarrhea (18%), epistaxis (12%), and hypertension (14%). The most common AEs in the placebo group were diarrhea (13%), nausea (13%), headache (13%), hypertension (13%), and thrombocytopenia (13%).

In the FIT1 trial, 16% of patients in the fostamatinib group had at least a single SAE (febrile neutropenia, immune thrombocytopenic purpura, thrombocytopenia, retinal tear, diarrhea, pneumonia, syncope, vaginal hemorrhage, or epistaxis) compared to 20% in the placebo group (anemia, congestive cardiac failure, gastrointestinal hemorrhage, menorrhagia, chronic obstructive pulmonary disease, epistaxis). In the FIT2 trial, 10% of patients in the fostamatinib group had at least a single SAE (epistaxis, bronchitis, contusion, decreased platelet count, plasma cell myeloma, transient ischemic attack, or hypertensive crisis) compared to 26% in the placebo group (thrombocytopenia, menorrhagia, muscle rupture, infection, or petechiae). In the FIT1 trial, 16% of patients in the fostamatinib group withdrew due to any AE (abdominal pain, diarrhea, neutropenia, thrombocytopenia, increased ALT, chest pain, pneumonia, or syncope) compared to 8% in the placebo group (abdominal discomfort or epistaxis). In the FIT2 trial, 4% of patients in the fostamatinib group withdrew due to any AE compared to 9% in the placebo group. In the fostamatinib group, 1 patient (2%) withdrew due to plasma cell myeloma and 1 due to headache. In the placebo group, 1 patient (4%) withdrew due to diarrhea and 1 due to hypertension. One patient in the FIT1 trial died in the placebo group due to sepsis. In the FIT2 trial, a single patient died in the fostamatinib group due to plasma cell myeloma.

In the FIT1 trial, |||||| of patients in the fostamatinib group experienced an infection compared to |||||| in the placebo group. In the FIT2 trial, |||||| of patients in the fostamatinib group and |||||| of patients in the placebo group experienced an infection. In both the FIT1 and FIT2 trials, ||||| of patients in the fostamatinib group experienced neutropenia compared to ||||| in the placebo group. In the FIT1 trial, |||||| of patients in the fostamatinib group experienced elevated liver transaminase levels compared to ||||| in the placebo group. In the FIT2 trial, ||||| of patients in the fostamatinib group experienced elevated liver transaminase levels compared to ||||| in the placebo group.

Critical Appraisal

The FIT1 and FIT2 trials were at an overall low risk of bias, although there were some concerns regarding selective outcome reporting (sensitivity analyses and subgroup analyses were not pre-specified) and a potential for unblinding due to high dropout rates due to a lack of response. In both the FIT1 and FIT2 trials, the fostamatinib and placebo groups were generally balanced in baseline characteristics, although some baseline imbalances in each trial may have introduced bias. For example, there were differences in the rates of specific previous ITP treatments used in the FIT1 trials and differences in the rate of prior splenectomy in the FIT2 trial, as well as higher concomitant steroid use in the placebo group in both trials. The rate of study discontinuation was high in both the FIT1 and FIT2 trials and was imbalanced between study groups in both trials, primarily due to discontinuation from the trials because of a lack of treatment response. Because patients discontinuing due to a lack of response were treated as nonresponders and an intention-to-treat (ITT) approach to the analysis was used, the high discontinuation rate did not appear to bias the primary outcome. However, for the SF-36 outcome, the high study-discontinuation rate meant limited data were available at week 24 (e.g., 1 patient in the placebo group at week 24 in the FIT1 trial, and 2 patients in the placebo group at week 24 in the FIT2 trial). It was therefore not possible to draw any meaningful conclusions from the SF-36 data at week 24 due to the limited amount of data from study discontinuations. Given the small number of patients in each subgroup and low event rates, there was likely insufficient power to detect any differences between treatments in these subgroups. This is reflected by wide CIs in the RD.

The small number of patients and low event rates for certain outcomes (ITP bleeding scale [IBLS] and WHO bleeding scale scores, bleeding-related SAEs, and the use of rescue therapy) make it challenging to draw conclusions about any difference between treatment groups for these secondary end points. These outcomes may also have been biased by imbalances in concomitant steroid use. Neither the FIT1 nor the FIT2 trial were powered for secondary end points and there was no adjustment for multiplicity for secondary end points; these outcomes should therefore be interpreted with caution.

Clinical experts indicated that the population of the FIT1 and FIT2 trials are broadly comparable to the population of patients with ITP in Canada, and the results of these trials are therefore likely generalizable to the Canadian population. The long duration of ITP and multiple previous treatments among patients in the FIT2 trial mirrors what is commonly seen among patients with ITP in clinical practice in Canada. However, the clinical experts did note generalizability concerns with the FIT1 and FIT2 trials in some Canadian contexts, as patients in both trials were predominantly White. The experts also noted that, because patients with secondary ITP were excluded from the FIT1 and FIT2 trials, findings may not be generalizable to those with secondary ITP. Further, the specific types of previous treatments used in the FIT1 and FIT2 trials differ from those commonly seen at a similar point in ITP treatment in Canada. The clinical experts pointed out that, based on the duration of ITP for patients in the FIT1 and FIT2 trials, a greater portion of chronic ITP patients in Canada would have had a prior splenectomy. Moreover, the extent of previous rituximab use in the FIT1 trial is higher than what would be seen in Canada at a similar stage of treatment. In terms of outcome assessment in the FIT1 and FIT2 trials, the clinical experts noted that bleeding outcomes are likely the most important in practice, but that treatment response is most commonly measured by platelet counts.

The FIT1 and FIT2 trials provided limited data on clinically important outcomes such as HRQoL, rescue therapy, and bleeding events. The clinical experts do not use the IBLS and WHO bleeding scale in practice, and the relevance of the bleeding outcome scales used in the trials is unclear. Further, the event rates for the post hoc bleeding-related SAE outcome made it challenging for the clinical experts to comment on the relevance or meaningfulness of these findings. The clinical experts found that the lower rate of rescue therapy use among patients treated with fostamatinib compared to placebo in the FIT1 and FIT2 trials could be meaningful, but they also noted the relatively low event rates. Similarly, given the small numbers and low event rates in the subgroup analyses, the clinical experts could not draw meaningful conclusions about whether any subgroup differences were likely to exist (e.g., based on a prior splenectomy or TPO-RA treatment). Another challenge with both the FIT1 and FIT2 trials is that the comparator is placebo. In chronic ITP, if the platelet count is below 20,000/µL, as it was at baseline for patients in both trials, the clinical experts (and clinical practice guidelines) indicated that treatment would be warranted. Placebo may therefore not be an appropriate comparator for fostamatinib. Indeed, the FIT1 and FIT2 trials do not address the comparative efficacy of fostamatinib against other second- or third-line ITP treatments.

Indirect Comparisons

Description of Studies

Two indirect treatment comparison (ITC) studies were reviewed. The Wojciechowski et al.⁷ study was a systematic review and ITC comparing fostamatinib to 3 TPO-RAs (avatrombopag, eltrombopag, and romiplostim) among patients with chronic ITP who had inadequate response to previous therapy. The sponsor also submitted a systematic review and ITC⁸ in which fostamatinib was compared to rituximab among patients with chronic or persistent ITP.

In the Wojciechowski study, the authors conducted a network meta-analysis (NMA) using a Bayesian framework, Markov chain Monte Carlo method, and fixed-effects model. The authors assessed consistency using a modified Bucher approach and used trace plots to assess convergence. They assessed the following outcomes: durable platelet response, need for rescue therapy, and WHO bleeding events, all up to 24 weeks. No sensitivity analyses were conducted. The authors reported outcomes as odds ratios (ORs) with 95% credible intervals (CrIs). Seven phase III, double-blind RCTs were included and contributed data on various clinical outcomes for the ITC: 2 trials for avatrombopag, 1 for eltrombopag, 2 for romiplostim, and 2 for fostamatinib. Six of the trials were placebo-controlled, and 1 compared avatrombopag with eltrombopag. In this ITC, the number of patients ranged from 12 to 135 in active treatment arms, the length of follow-up ranged from 24 to 36 weeks, and the median duration of disease ranged from 1.6 to 10.8 years. The median ages and median platelet counts at baseline ranged from 41 to 57 years and 14 × 10⁹/L to 24 × 10⁹/L, respectively. The number of previous treatments and concomitant ITP therapy varied across trials as well. The definitions of durable response were relatively similar for all the treatments to allow for comparisons, although the definitions were different in the FIT 1 and FIT 2 trials (4 of 6 visits over weeks 14 to 24) compared with the other trials (at least 6 of the last 8 weeks of treatment). To compare bleeding events among all treatments, it was assumed that WHO grade 2 to grade 4 bleeds were equivalent to grade 2 to grade 5 bleeds reported in NCT00102336 and moderate to severe bleeds reported in the FIT1 and FIT2 studies.

In the sponsor-submitted ITC, a random-effects model was used; there was no assessment of consistency and convergence was assessed via the Gelman-Rubin statistic. One outcome, overall platelet response, was assessed in this ITC and reported as ORs, which represent the relative likelihood of achieving a platelet response when receiving 1 therapy compared against another therapy. The results were presented using the posterior median treatment effects and 95% CrIs). The authors performed a sensitivity analysis using different definitions of platelet response and doses of rituximab. Six RCTs were included and contributed evidence. Authors of the ITC indicated that the inclusion of fostamatinib in the treatment paradigms aims to address patients with the greatest unmet need. Patients who are successfully treated with a splenectomy or TPO-RAs are unlikely to require additional treatment with a new intervention. To address the patients with the greatest unmet need, fostamatinib focuses on patients who do not have access to long-term, effective therapy options, including those who receive short courses of rituximab and those on a watch-and-rescue regimen. In the trials included in the ITC, the number of enrolled patients ranged from 57 to 138. The trial duration ranged from 4 weeks to 78 weeks. Three doses of rituximab were evaluated: 2 or 4 once-weekly 375 mg/m² doses, 2 once-weekly doses of 750 mg/m², or 4 once-weekly doses of 100 mg/m². The definition of platelet response varied across the included trials.

Efficacy Results

In the Wojciechowski study, based on evidence from 6 studies, the results of the NMA suggested that no treatment was favoured when fostamatinib was compared with various TPO-RAs for durable platelet response. Data on reduction in the use of concomitant ITP therapies were not available for fostamatinib. Based on 6 studies, the NMA suggested that no treatment was favoured when fostamatinib was compared with TPO-RAs for need for rescue therapy. Results from 7 studies suggested that no treatment was favoured when fostamatinib was compared with TOP-RAs for the incidence of any bleeding events. Further, based on 6 studies, the NMA suggested that no treatment was favoured when fostamatinib was compared with TPO-RAs for the incidence of WHO grade 2 to grade 4 bleeding events.

In the sponsor-submitted ITC, results demonstrated that fostamatinib was favoured compared to placebo for the outcome of overall platelet response (OR = 4.85; 95% CrI, 1.86 to 14.45). Fostamatinib was also favoured when compared to rituximab for the outcome of overall platelet response (OR = 4.93; 95% CrI, 1.44 to 18.93).

Harms Results

Only the Wojciechowski study assessed AEs. Based on the results from 5 studies, the NMA suggested that no treatment was favoured when fostamatinib was compared with TPO-RAs for the incidence of any AEs.

Critical Appraisal

The Wojciechowski ITC did not discuss how any potential biases in the trials could have an impact on data analyses in the ITC and the possible solutions. For example, it did not discuss whether sensitivity analyses were conducted to assess the impact of studies with poor quality. Multiple clinical outcomes, including the incidence of AEs, were evaluated in this study, which allows for a comprehensive evaluation of the clinical benefits and risks of the study drugs. Definitions of these outcomes were similar across the trials. Trial characteristics and patients’ baseline characteristics in the studies included in the systematic review and ITC were reported. Potential sources of heterogeneity with respect to the baseline characteristics, such as disease duration (which ranged from 1.6 to 10.8 years), number of previous treatments, and concomitant ITP medication, were identified based on these data. This difference in patient baseline characteristics may vary the response between groups and may not allow groups to be comparable. The analysis of efficacy and safety data presented was limited by the size of the evidence base. Due to the small evidence base and potential heterogeneity across all studies, the results of this analysis are largely noninformative due to imprecision and bias.

Only overall platelet response was evaluated in the sponsor-submitted ITC. It is unclear whether treatment with fostamatinib would be useful in improving clinical outcomes, such as reduction in subsequent bleeding events and the need for rescue therapy, and improvement in patients’ HRQoL. Potential sources of heterogeneity with respect to the baseline characteristics were identified in this ITC. During the feasibility analysis period, several potential treatment-effect modifiers, such as baseline demographic characteristics, medical history (e.g., time since ITP diagnosis, prior treatment for ITP, and concomitant medications), were identified by the sponsor. However, the clinical experts consulted by the sponsor indicated that none of these patient characteristics could be considered treatment-effect modifiers in the study population (patients with chronic and persistent ITP who can receive fostamatinib or rituximab) due to a lack of evidence. Other patient characteristics in the study population could be treatment-effect modifiers that were not measured. Heterogeneity across the included trials therefore needs to be further assessed and adjusted. Rituximab was the only comparator in this ITC, and its study results can only be generalized to patients with persistent or chronic ITP who did not receive prior TPO-RAs therapy or had not undergone a splenectomy. Indeed, the lack of comparators was a major limitation of both the sponsor-submitted and the Wojciechowski ITCs.

Other Relevant Evidence

Description of Studies

The FIT3 trial was considered other relevant evidence.⁹ This was an open-label extension study of the FIT1 and FIT2 trials to examine the efficacy and safety of long-term fostamatinib among patients with chronic or persistent ITP at 54 sites in 16 countries (Canada, the US, Australia, the European Union, and the UK). The trial consisted of monthly visits for 18 months, then every-other-month visits for a maximum of 5 years of treatment. A total of 123 patients from the FIT1 and FIT2 trials completing the week 24 evaluation or withdrawing early (starting at week 12) due to a lack of response were eligible for this trial. All patients received open-label fostamatinib. Patients were assigned to 1 of 2 treatment groups, responders and nonresponders, depending on their response in the FIT1 or FIT2 trial. The responders group (last platelet count ≥ 50,000/µL) initiated open-label fostamatinib treatment using the dosage and regimen (150 mg twice a day or 100 mg twice a day) that achieved a stable platelet count in the previous study, whereas the nonresponders group (last platelet count < 50,000/µL) initiated their treatment with 100 mg twice a day during the trial (Figure 5). At month 1, the dosage for patients with a platelet count of less than 50,000/µL and tolerating the study drug well was increased to 150 mg twice a day. However, the dosage of fostamatinib was reduced to as low as 100 mg once a day if any dose-limiting AEs were observed among patients. The primary efficacy outcome was achievement of a platelet response by 12 weeks and maintenance for 12 months. A stable platelet response was a platelet count of at least 50,000/μL at 4 or more of 6 biweekly visits during weeks 14 to 24 or, for patients initiating fostamatinib in the extension phase, at least 1 platelet count of 50,000/μL or greater in the first 3 months followed by platelet counts of at least 50,000/μL at the subsequent 2 of 3 monthly visits without the use of rescue medication. The primary efficacy outcome had 2 versions. For version 1, efficacy was assessed among patients who were on active treatment in either the FIT1 or FIT2 trial, in the current extension study, or in both a FIT trial and the extension trial. For version 2, efficacy was assessed among patients assigned to placebo in either of the prior FIT1 or FIT2 trials. The secondary efficacy outcomes were reported as the duration of platelet response among patients and the response among patients with a reduction in the dose of concomitant ITP medication while maintaining an adequate platelet count. For the safety measurement, the outcomes assessed and summarized in the report were: the frequency and severity of bleeding according to the IBLS and WHO bleeding scale; change from baseline in liver function, blood pressure, and neutrophil count; and the incidence and severity of gastrointestinal effects, infection, and overall AEs. In the FIT3 trial, 60% of patients were female, the median age was 52 years (SD = 16), and patients were predominantly White (92%). A total of 59 patients were from the FIT1 trial and 64 patients were from the FIT2 trial.

A post hoc analysis of the FIT1, FIT2, and FIT3 trials by Boccia et al. (2020) was also considered relevant evidence. Authors compared the platelet response rate (≥ 50,000/µL and ≥ 30,000/µL at any visit, without receiving rescue therapy within 4 weeks) in patients who received fostamatinib as second-line therapy to those who had received fostamatinib as third- or later-line therapy for chronic ITP. A total of 145 patients were included, 32 receiving fostamatinib (median age of 50 years, 59% female) and 113 as later-line treatment (median age of 54 years, 60% female).

Efficacy Results

For the primary efficacy outcome (version 1) in the FIT3 trial, 19 patients (15.4%) had a platelet response within 12 weeks of taking fostamatinib and maintained a stable platelet response for at least 12 months after achieving the initial response (95% CI, 9.6 to 23.1). For the primary efficacy outcome (version 2), among 44 patients who were treated with placebo in the FIT1 or FIT2 trial and fostamatinib in the FIT3 trial, 10 (22.7%) were responders, while 34 (77.3%) remained nonresponders in both prior trials and in the FIT3 trial.

In the post hoc analysis by Boccia et al., 25 patients (78%) receiving fostamatinib as second-line therapy achieved a platelet response of a least 50,000/µL compared to 54 patients (48%) on later-line therapy. The authors reported that the response decreased with each additional line of prior therapy.

Harms Results

Most patients (80%) experienced at least 1 AE during the treatment phase of the FIT3 trial. The most frequently reported AEs were diarrhea (29%), hypertension (18%), petechiae (15%), epistaxis (15%), headache (12%), upper respiratory tract infection (11%), dizziness (11%), contusion (10%), nausea (9%), vomiting (9%), fatigue (8%), cough (8%), and thrombocytopenia (8%). Serious AEs with were reported for 28% of patients, with thrombocytopenia being the most frequently reported among 7% patients. A total of 18 patients (15%) withdrew from the trial due to an AE and 4 people (3%) died.

In the post hoc analysis by Boccia et al., the authors reported that AE rates were 72% for second-line therapy and 94% for later-line therapy. The most common AEs were hypertension (31% in the second line versus 19% in the later line), diarrhea (25% versus 39%, respectively), upper respiratory tract infections (16% versus 11% respectively), and elevated liver transaminase levels (26% versus 16%, respectively).

Critical Appraisal

The main limitation of the FIT3 trial is the open-label design and the lack of a comparator group. The open-label design may influence the perception of improvement by patients and clinicians, potentially affecting the reporting of harms and efficacy measures. Additionally, there was a potential for survival and selection bias as the other 13 patients who discontinued the prior studies due to AEs were excluded. This could result in a greater enrolment of patients who were better able to tolerate fostamatinib and possibly fewer AEs being reported. The FIT3 trial also saw a high rate of discontinuation (76.4%) during the open-label phase. The limitations with the study design make it challenging to interpret the results and form conclusions on long-term efficacy and safety.

As participants in the FIT3 trial were predominantly White (92%), the results may not be generalizable to other racial groups commonly seen at some centres in Canada. The experts also noted that patients with secondary ITP were excluded from the FIT3 trial, and trial findings therefore may not be generalizable to those with secondary ITP. The clinical experts also noted that the co-interventions (i.e., concomitant ITP medication) used in the FIT3 trial, as with the FIT1 and FIT2 trials, reflect real-world practice in Canada. Similar to the FIT1 and FIT2 trials, the FIT3 trial provided limited data on clinically important outcomes such as quality of life, rescue therapy, and bleeding events.

Internal and external validity concerns from the FIT1, FIT2, and FIT3 trials apply to the post hoc analysis by Boccia et al., which was subject to concerns related to selective outcome reporting. The analysis was not pre-planned, and the outcome used differs from those in the FIT1 and FIT2 trials as response to therapy was based on the platelet count at a single visit. Patients in the different treatment groups were not randomized and there was no adjustment for confounding. As such, concerns related to selection bias and bias due to confounding reduce the certainty in these results.

Conclusions

Management of chronic ITP is challenging as patients frequently relapse or are refractory to treatments and therefore often cycle through multiple ITP treatments. Treatment is complicated by a lack of evidence on comparative efficacy and safety of second- and subsequent-line treatment options, access issues, and safety and/or tolerability of available options. In 2 double-blind RCTs, fostamatinib, which is an ITP treatment with a novel mechanism of action, led to a modest improvement in platelet count response compared to placebo among patients with heavily pre-treated, primary, chronic ITP. There were limited or no data on outcomes important to patients such as bleeding rates, symptoms, and quality of life. The impact of fostamatinib on these outcomes therefore remains unclear. Subgroup analyses (based, for example, on previous lines of therapy) were not able to provide insight into which patient groups are most likely to respond to treatment. It is also difficult to draw conclusions about the comparative efficacy of fostamatinib versus other ITP treatments. Two ITC studies were included in this review, suggesting that fostamatinib may be comparable to TPO-RAs and had favourable efficacy to rituximab in terms of platelet count response. However, these studies have important limitations, and it is challenging to draw firm conclusions about comparative efficacy based on their results. In the FIT1 and FIT2 trials, fostamatinib appeared to lead to a higher rate of adverse effects, such as diarrhea, nausea, hypertension, and elevated liver transaminase compared to placebo, while the FIT3 trial did not identify any long-term safety concerns beyond these adverse effects. Overall, this review suggests that fostamatinib is another potential treatment option for patients with chronic, heavily pre-treated primary ITP. The drug leads to a platelet count response in a modest proportion of patients and is generally well tolerated compared to placebo, although its comparative efficacy and safety versus other ITP treatments, and its effect on patient-important clinical outcomes, remain unclear.

Introduction

Disease Background

Immune thrombocytopenia is a “primary” or “secondary” autoimmune disorder characterized by low platelet counts and increased bleeding risk.¹ It is thought to be caused by antibodies directed against platelet antigens, leading to increased platelet destruction.¹ Primary ITP is not triggered by a specific condition or event while secondary ITP is caused by or associated with another condition, such as chronic lymphocytic leukemia, systemic lupus erythematous, antiphospholipid syndrome, among others.¹ Primary ITP accounts for approximately 80% of cases.¹⁰ It is also defined based on the duration, with acute or newly diagnosed ITP referring to the first 3 months after diagnosis, persistent ITP referring to 3 to 12 months after diagnosis, and chronic ITP referring to more than 12 months after diagnosis.¹ Little contemporary data are available on the incidence and prevalence of ITP in Canada. A 2010 narrative review of international studies suggested that the incidence of ITP among adults is approximately 3.3 per 100,000 per year, while the prevalence is 10 per 100,000, both of which increase with increasing age.¹¹ An American study using data from 2010 to 2016 suggested that the annual incidence of ITP in the US was 6.1 per 100,000 persons.¹² The rate of fatal hemorrhage among patients with ITP was been estimated to be between 0.016 and 0.039 cases per patient-year, and this rate increases with age.¹³ The predicted 5-year mortality rate for patients 60 years of age and older was 48% in a study of 1,817 patients with ITP.¹³ Further, the authors of this study estimated that a 30-year-old woman with ITP would lose 15 quality-adjusted life-years from her life expectancy.¹³

Patients with ITP may be asymptomatic, although patients can experience bleeding and other symptoms.¹⁴ Bleeding can be mild; for example, patients may experience petechiae, purpura, or nosebleeds.¹⁴ In cases of intracranial hemorrhage or gastrointestinal bleeding, it can be more severe or critical.¹ Indeed, severe or critical bleeding is a major concern among patients with ITP. Predictors of critical bleeding include platelet count (< 10,000/µL or < 20,000/µL), previous bleeding, and chronic ITP (> 12 months in duration). Patients with ITP also commonly experience fatigue.¹⁵ Patients with ITP have a reduced quality of life, resulting from fatigue, bleeding, and ITP treatments.¹⁶

Because ITP is considered a diagnosis of exclusion, the diagnostic evaluation primarily concerns excluding other possible causes of low platelet count and/or finding potential conditions leading to low platelet counts (i.e., secondary ITP).¹ In addition, because ITP is an isolated thrombocytopenia, patients with the disease do not have anemia or leukopenia.¹ Diagnosis involves taking a history (questioning regarding bleeding and symptoms), physical examination, and laboratory testing (e.g., a complete blood count and peripheral blood smear).¹ Clinical experts suggested that initial diagnosis and management may be carried out by internal medicine clinicians, while patients with chronic ITP will generally be managed by hematologists.

Standards of Therapy

The need for treatment to increase platelet counts among patients with ITP is based on assessments of bleeding (site, acuity, and severity), platelet count, bleeding risk factors, and previous treatments. Treatment to increase platelet count is generally recommended if the platelet count is below 20,000/µL to 30,000/µL and/or if the patient is experiencing bleeding. Patients with severe or critical bleeding are recommended to receive urgent treatment to stop bleeding and raise platelet counts.

The main goals of therapy in ITP are to prevent severe or critical bleeding, reduce or eliminate patients’ symptoms, minimize adverse effects from treatments, and ultimately improve patient quality of life.² Treatments are recommended to increase platelet levels to above 20,000/µL to 30,000/µL, which appears to reduce the risk of major bleeding.^2,3 American and International guidelines recommend that, for initial treatment of newly diagnosed ITP, corticosteroids (for 2 weeks then tapered) or IVIG (for 1 to 5 days) be used as first-line therapy. Anti-D immune globulin is another alternative in patients with bleeding or at high risk of bleeding.^2,3 Long-term corticosteroid treatment is generally not a recommended treatment option as the harms outweigh the benefits.²

After corticosteroids or IVIG are stopped, many patients (one-third of patients in the first year and up to 80% within 5 years) experience a relapse in their condition in the form of reduced platelet counts and/or symptoms.² Once patients have relapsed, the sequence of subsequent treatments is less clear.^2,3 Multiple second- and third-line treatments are available for ITP; however, there is a lack of comparative efficacy data to provide evidence on a clear sequential treatment pathway. Possible treatment options include a splenectomy; rituximab; TPO-RAs such as romiplostim or eltrombopag; fostamatinib; and immunosuppressants (e.g., azathioprine and cyclophosphamide). The International Consensus Report on the Investigation and Management of Primary ITP highlights these treatment options in subsequent-line treatment of ITP but does not identify a preferred pathway among them.² These guidelines state that the recommended option is generally based on available resources and patient preferences. The guidelines further note that “robust” evidence supports the use of TPO-RAs, rituximab, and fostamatinib in subsequent-line treatment of ITP. The American Society of Hematology guidelines for ITP suggest rituximab, a splenectomy, or TPO-RAs as second-line treatment options.³ They state that the decision is based on patient preferences and other patient-specific factors (age and comorbidities), as well as access (cost and availability). These guidelines also acknowledge the low certainty of the evidence on comparative efficacy and state that individualization of therapy and shared decision-making (based on these factors) are important in identifying the appropriate subsequent-line ITP treatment.

In the Canadian context, the choice of subsequent-line treatment depends on patient-specific factors (e.g., increased susceptibility to adverse effects of a treatment, contraindications, preferences) as well as access (i.e., whether a treatment is listed on a provincial drug formulary and/or whether the patient meets the criteria for reimbursement). Consequently, some options may not be available or appropriate to all patients. Further, some patients may not be surgical candidates and splenectomy carries risks such as infection. Rituximab has been shown to be effective in achieving a platelet response but has a risk of fatal infusion reactions, hepatitis B reactivation, and progressive multifocal leukoencephalopathy, among other harms.² Indeed, many of the second- and third-line treatment options carry risks of important harms. For example, TPO-RAs increase the risk of bone marrow reticulin fibrosis and arterial and venous thrombosis.² Options also differ in terms of their administration. For example, rituximab is given as an infusion at a clinic or hospital over several weeks, while eltrombopag is a daily continuous oral medication that cannot be taken within several hours of calcium. Overall, the potential chance of achieving platelet response must be considered against the potential harms of the different agents and administration factors, as well as access issues.³

The clinical experts consulted by CADTH emphasized that not all patients respond to treatment with second- or third-line treatment options. Further, patients become refractory to treatment options or relapse after achieving remission. Chronic ITP is therefore characterized by a chronic relapsing course and multiple lines of therapy over time.

Drug

Fostamatinib is indicated for the treatment of chronic ITP in adult patients who have had an insufficient response to other treatments.⁴ Fostamatinib treatment should be initiated and remain under the supervision of a physician who is experienced in the treatment of hematological diseases.⁴ Fostamatinib reduces the destruction of platelets via inhibition of spleen tyrosine kinase.⁴ Fostamatinib is initiated at a dosage of 100 mg twice daily and is taken orally. If the platelet count has not increased to at least 50,000/µL after 4 weeks, then the dosage can be increased to 150 mg twice daily.⁴ Fostamatinib underwent an expedited review at Health Canada. The sponsor is requesting reimbursement for fostamatinib for the treatment of thrombocytopenia in adult patients with ITP who have had an insufficient response to a TPO-RA in jurisdictions where TPO-RA reimbursement is available, or after failure of corticosteroids and other earlier-line treatments in jurisdictions where TPO-RA reimbursement is not available.

Fostamatinib is contraindicated in patients who are hypersensitive to the drug or any ingredient in the formulation, and is also contraindicated in pregnancy.⁴ The product monograph carries warnings for bone remodelling, stating that fostamatinib targets pathways involved in bone metabolism.⁴ The monograph notes that the effects of fostamatinib on bone remodelling are unclear; however, there are potential risks in those with actively growing bones (e.g., children and young adults), and patients with osteoporosis or fractures should be closely monitored. Hypertension has been reported among patients treated with fostamatinib.⁴ Patients with hypertension may be more susceptible to hypertensive effects, and blood pressure should be closely monitored.⁴ Other warnings in the monograph include heart rate and conduction abnormalities, gastrointestinal side effects (particularly diarrhea), neutropenia, elevated transaminases, and infections.⁴

Stakeholder Perspectives

Patient Group Input

This section was prepared by CADTH staff based on the input provided by patient groups.

One patient group submission, authored by the PDSA, was received for this review. The PDSA, founded in 1998, is a US-based international nonprofit organization and a registered nonprofit corporation in Canada. Its members are dedicated to enhancing the lives of patients with ITP and other platelet disorders through advocacy, education, research, and support. The PDSA has 635 Canadian adults and children in its database and has 7 support groups, including chapters in London, Niagara, Toronto, Waterloo, Ottawa, and Vancouver. The PDSA has received funding from argenx, Amgen, Dova/Sobi, Novartis, UCB, CSL Behring, Principa, Pfizer, Sanofi, Momenta, and Rigel. The association did not receive help from outside the patient group to prepare its submission. The submission includes patient comments from the PDSA Facebook page gathered from the US and Canada between 2018 and present — these comments were from people who had taken fostamatinib and therefore pertain only to the section on patient experiences with fostamatinib. The submission did not describe how the rest of the input was gathered.

Disease Experience

Immune thrombocytopenia is unpredictable and affects both the patient and their entire family. Patient quality of life is affected in multiple ways. Patients are fearful about the risk of life-threatening bleeding but also face physical and emotional consequences, such as fatigue, anxiety, depression, pain, and sleep disturbances. These symptoms are often more concerning to patients than platelet counts. Patients with ITP may restrict or avoid activities (e.g., travelling or participating in sports), and require frequent monitoring, which, along with symptoms, interferes with their daily activities. The disease can also make treatment, including medical procedures and surgery, more complex. These factors further lead to anxiety, fear, and depression.

Experience With Treatment

Several therapies are available for treatment of ITP, and each have different risk-benefit profiles and limitations. Multiple therapies can also be used at once. Prednisone can increase platelet counts but is recommended for short-term use due to the risk of side effects with longer-term use. Both IVIG and anti-D immune globulin are short-term treatment options that can be used as rescue therapy to increase platelet counts but are not suitable for long-term therapy. Treatments aimed at producing long-term increases in platelet counts include rituximab, splenectomy, TPO-RAs, and fostamatinib. However, not all patients will respond to these therapies, and side effects are a concern. Patients therefore often cycle through different options to find a therapy that will be tolerated and raise platelet counts.

The PDSA identified a selection of Facebook posts among people with lived experience of fostamatinib treatment from 2018 to present. Patients that commented had generally tried other therapies. Patients commenting suggested that fostamatinib was effective in increasing platelet counts, with some noting that it took approximately 2 weeks to increase platelet counts. Some patients who had been on fostamatinib long-term (16 months to 2 years) stated that they continued to have a platelet count response. Some patients commented that they had not experienced adverse effects, while others noted adverse but manageable effects, such as diarrhea, elevated blood pressure, or stomach upset.

Improved Outcomes

It is difficult to predict who will respond to a particular treatment and who will develop resistance to a treatment over time. Further, patients may not be able to afford or access available options. It is therefore important that patients have options available in case they do not respond to a therapy, the therapy stops working, or they experience bleeding. Patients prefer treatments that do not affect their daily lives and patients find it easier to take a pill than to go to a hospital or clinic to receive treatment. Patients also prefer a treatment that has minimal side effects and a durable response. Patients experience anxiety from possible bleeding, as well as nose bleeds, mouth blisters, and fatigue, and want a therapy that means they will not live in fear of when their next bleed will be. An ITP therapy should improve quality of life, not reduce it.

Clinician Input

Input From Clinical Experts Consulted by CADTH

All CADTH review teams include at least 1 clinical specialist with expertise in the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., providing guidance on the development of the review protocol; assisting in the critical appraisal of clinical evidence; interpreting the clinical relevance of the results; and providing guidance on the potential place in therapy). In addition, as part of the fostamatinib review, a panel of 4 clinical experts from across Canada was convened to characterize unmet therapeutic needs, assist in identifying and communicating situations where there are gaps in the evidence that could be addressed through the collection of additional data, promote the early identification of potential implementation challenges, gain further insight into the clinical management of patients living with a condition, and explore the potential place in therapy of the drug (e.g., potential reimbursement conditions). A summary of this panel discussion follows.

Current Treatments

Standard first-line therapy for ITP includes corticosteroids, and IVIG (or, in rhesus disease–positive patients, rhesus-immune globulin) is often added when an immediate increase in platelets is required, although its effect is often transient. A significant proportion of patients will not respond to steroids and, of those that do, many will relapse once steroids are tapered. At this point, traditional second-line therapy is a splenectomy if the patients is a suitable candidate. More recently, rituximab has emerged as an alternative second-line therapy. If both a splenectomy and rituximab have failed (or are contraindicated), a large number of third-line therapies are available, including immunosuppressant medications such as azathioprine or cyclophosphamide, or TPO-RAs such as eltrombopag or romiplostim. There is little evidence to guide the selection of third-line therapy, and decisions depend on local reimbursement considerations as patient-specific factors.

Treatment Goals

Broad clinician treatment goals are to reduce bleeding and prolong life. Increasing the platelet count is generally considered to be a reasonable surrogate for those 2 goals. Improving quality of life is also an important goal but must be balanced against the inconvenience and side effects of the treatments used (e.g., fatigue, cognition, mood, interference with daily life, and frequent hospital visits), which many clinicians may overlook in their focus on the patient’s platelet count.

Unmet Needs

The current treatment paradigm for ITP poses myriad challenges. Not all patients respond to available therapies, and even if remission is achieved, long-term remission is not guaranteed. Durable remission for ITP remains a challenge. Further, while corticosteroids and IVIG are generally accessible to patients, accessibility to appropriate second- and third-line therapies can be a challenge. This is because not all options are reimbursed in each province or because reimbursement criteria differ across provinces. For example, in Ontario patients must fail 2 or more therapies after steroids and IVIG before being eligible for TPO-RAs, meaning that these agents are not available to many patients until later in the treatment pathway. Administration of existing therapies can also be a challenge, for example, when there is a need to travel to a hospital or clinic for administration of rituximab. Adhering to oral TPO-RA dosing regimens can also be difficult as the drugs must be administered on an empty stomach. There are also adverse effects with existing treatments — a splenectomy carries short-term perioperative risks as well as longer-term risks of thrombosis and infections with encapsulated bacteria, while rituximab increases susceptibility to hepatitis B reactivation and increases vulnerability to opportunistic infections. The availability of therapies with demonstrated efficacy, convenience of administration, and a low risk of adverse effects would therefore fill an unmet need for treatment of ITP.

Place in Therapy

Contemporary ITP guidelines suggest that, in general, a splenectomy or rituximab can be considered second-line therapy. Several third-line options are available; however, the comparative efficacy of these drugs is unclear, it can be difficult to know what the best treatment option is for a particular patient, and there is often no single clearly defined treatment pathway. Decisions end up being driven largely by access. Given the lack of comparative efficacy data, the influence of patient-specific factors on decisions, and the current reimbursement landscape, it is challenging to identify the optimal place in the therapeutic algorithm for fostamatinib, which is a novel spleen tyrosine kinase inhibitor recently approved for the treatment of ITP. The clinical experts consulted by CADTH noted that rituximab or a splenectomy are reasonable second-line choices (TPO-RAs may also be considered second-line choices in some patients). The safety profile of fostamatinib and the fact that it is administered orally suggest it may be considered a reasonable third-line therapy rather than reserved for patients who have failed or do not have access to TPO-RAs, as has been proposed by the sponsor. However, regardless of where it sits in the therapeutic algorithm, it would be advantageous for clinicians to have fostamatinib as an additional treatment option for specific patients.

Patient Population

Subgroup analysis of data from randomized controlled trials (RCTs) suggests that patients who have failed fewer prior ITP treatments may respond to fostamatinib better than those who have been more heavily pre-treated. As these data are prone to selection bias, the clinical experts stated that they would not base their treatment decisions on this finding and noted that it would be helpful to have an additional treatment option even in patients who had failed many previous therapies. The ITP population is heterogenous, and the available data and current understanding of ITP pathophysiology make it impossible to determine which specific patients will respond best and who are most susceptible to adverse effects. However, the panel agreed that having fostamatinib as an option for patients would be desirable, regardless of where patients are in their disease course.

Assessing Response to Treatment

Bleeding is an important outcome in the treatment of ITP, and ultimately any treatment should reduce the occurrence of clinically important bleeding while improving quality of life. In practice, clinicians rely on platelet response, which is assumed to reduce the risk of clinically relevant bleeding and, as a secondary benefit, reduce the need for rescue therapy. No quality-of-life scales that are particular to ITP are used in practice. In general, an increase in platelet count can be seen as early as 2 weeks into treatment with fostamatinib, although some patients may not respond until week 12. If a response is observed, clinicians would likely continue to use the treatment long-term with monthly monitoring. A sustained response would generally be considered a platelet count of 30,000/µL to 50 000/µL for the duration of a treatment cycle (e.g., 24 weeks). If a response has not been seen by approximately 24 weeks, clinicians would generally consider that the treatment has not worked and would discontinue it. If there are issues related to safety or tolerability, treatment would generally be discontinued earlier, particularly if it is affecting a patient’s quality of life.

Prescribing Conditions

Clinicians practising general internal medicine frequently prescribe corticosteroids for the initial management of ITP. However, patients requiring second-line treatment are often referred to a hematologist. Patients with ITP for many years who have tried multiple therapies are often seen by multiple hematologists. While hematologists usually take responsibility for selecting treatment for ITP, primary care physicians may share responsibility for monitoring AEs.

Clinician Group Input

This section was prepared by CADTH staff based on input provided by clinician groups.

A group of 19 Canadian hematologists submitted input on fostamatinib. This group included clinicians from Alberta, Ontario, Nova Scotia, British Columbia, Quebec, and Newfoundland, and was informed by a literature review, current clinical practice guidelines, and clinical practice. The report also incorporated data from a survey conducted in July 2021 among 5 physicians based in the US with experience prescribing fostamatinib. The survey was conducted by Blue Ribbon Project Inc. and commissioned by Accelera Canada in partnership with Advocacy Solutions.

Current Treatments

The clinician group submission echoes the opinions of the expert panel. The clinician group noted that patients with ITP can have a wide range of clinical manifestations (from asymptomatic with low platelet counts to severe or life-threatening bleeding). Further, patients can follow a variable disease course, with sometimes long periods of stability and intermittent episodes of bleeding. Variability and unpredictability can be challenging and lead to a poor quality of life for patients as well as health-system impacts. Similar to the expert panel, the clinician group noted that many different treatment options are available, but patients often have relapses or do not respond to particular agents. Corticosteroids and/or IVIG are first-line options; however, there is a lack of data to guide the correct order or sequence of second- and subsequent-line therapies. In addition, the treatment paradigm differs across provinces in Canada due to differences in access and reimbursement policies.

Treatment Goals

Treatment goals highlighted by the clinician group align with what was reported by the clinical experts: patients want to alleviate symptoms, improve quality of life, and reduce the impact of ITP on their daily lives, in addition to reducing the risk of serious bleeding. Patients are also concerned about the side effects of medications and the costs of treatments. Clinician goals align with these patient goals, but clinicians also often focus on increasing platelet counts and minimizing health-system impacts. Finding an optimal treatment option that is tailored to the individual requires discussion with patients.

Unmet Needs

The clinician group submission echoed that of the clinical experts. The clinician group suggested that no treatments available to date have been shown to address key outcomes of interest for patients and patient goals of care, such as improvements in energy levels or mental health. The clinician group suggested that most treatments may actually worsen fatigue and mental stress. The submission suggested that there is a need for treatments that are better tolerated and that lead to better adherence, which can be a challenge in ITP. The clinician group stated that the patients with the greatest unmet need are those with severe refractory diseases who have failed first-line therapies and subsequent lines of therapy. The group added that having a drug that is given orally may improve adherence due its ease of administration.

Place in Therapy

The clinician group stated that fostamatinib is the first drug to target phagocytosis of platelets and therefore has a novel mechanism of action. The group suggested that fostamatinib would be likely be used after first-line therapy as second-line or subsequent-line therapy. The clinician group submission reported that fostamatinib would provide an alternative to other second- and subsequent-line therapies and should be considered before a splenectomy, immunosuppressive drugs, and rituximab and its biosimilars, and be comparable to maintenance treatments such as the TPO-RAs.

Patient Population

The clinician group submission stated that patients early in their ITP disease course may respond better to fostamatinib. Using it as a second-line therapy may offer advantages such as limiting exposure to complications or toxicities from other drugs. However, the greatest need is still in patients who have relapsed multiple times despite treatment.

Assessing Response to Treatment

The clinician group suggested that, in practice, platelet count and lack of rescue therapy were appropriate ways to assess response. The submission suggested that achieving the goals of the patient, clinician, and health care system would represent a clinically meaningful response. The clinician group stated that most effective response would be a prolonged life through the reduction of the risk of life-threatening bleeding. The submission from the clinician group suggested that an ITP treatment would be discontinued if there is disease progression (drop in platelets or increased bleeding), the patient develops side effects, or there is a need for rescue therapy.

Prescribing Conditions

The clinician group suggested that fostamatinib can be used in the outpatient setting (in hospital clinic) as well as in the emergency room and in hospitalized patients.

Drug Program Input

The drug programs provide input on each drug being reviewed through CADTH’s reimbursement review processes by identifying issues that may affect their ability to implement a recommendation. The implementation questions and corresponding responses from the clinical experts consulted by CADTH are summarized in Table 4.

Clinical Evidence

The clinical evidence included in the review of fostamatinib is presented in 3 sections. The first section, the systematic review, includes pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those studies that were selected according to an a priori protocol. The second section includes indirect evidence from the sponsor and indirect evidence selected from the literature that met the selection criteria specified in the review. The third section includes sponsor-submitted long-term extension studies and additional relevant studies that were considered to address important gaps in the evidence included in the systematic review.

Systematic Review (Pivotal and Protocol-Selected Studies)

Objectives

To perform a systematic review of the beneficial and harmful effects of fostamatinib for the treatment of thrombocytopenia in adult patients with of chronic ITP who have had an insufficient response to other treatments.

Methods

Studies selected for inclusion in the systematic review included pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those meeting the selection criteria presented in Table 5. Outcomes included in the CADTH review protocol reflect outcomes considered to be important to patients, clinicians, and drug plans.

The literature search for clinical studies was performed by an information specialist using a peer-reviewed search strategy according to the PRESS Peer Review of Electronic Search Strategies checklist.²¹

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946—) via Ovid and Embase (1974—) via Ovid. All Ovid searches were run simultaneously as a multi-file search. Duplicates were removed using Ovid deduplication for multi-file searches, followed by manual deduplication in EndNote. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concept was fostamatinib. Clinical trials registries searched included the US National Institutes of Health’s clinicaltrials. gov, the WHO International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

No filters were applied to limit retrieval by study type. Retrieval was not limited by publication date or by language. Conference abstracts were excluded from the search results. Appendix 1 provides detailed search strategies.

The initial search was completed on August 20, 2021. Regular alerts updated the search until the meeting of the CADTH Canadian Drug Expert Committee on December 15, 2021.

Grey literature (literature that is not commercially published) was identified by searching relevant websites from the Grey Matters: A Practical Tool For Searching Health-Related Grey Literature checklist.²² Included in this search were the websites of regulatory agencies (US FDA and European Medicines Agency). Google was used to search for additional internet-based materials. Appendix 1 provides more information on the grey literature search strategy.

Two CADTH clinical reviewers independently selected studies for inclusion in the review based on titles and abstracts, according to the predetermined protocol. Full-text articles of all citations considered potentially relevant by at least 1 reviewer were acquired. Reviewers independently made the final selection of studies to be included in the review, and differences were resolved through discussion.

In addition to the indirect evidence provided by the sponsor, additional indirect evidence that includes the patients, interventions, comparators, and outcomes specified in Table 5 was summarized and critically appraised, if considered relevant by CADTH. A focused literature search for NMAs dealing with immune thrombocytopenia was run in MEDLINE All (1946–) on August 20, 2021. No limits were applied to the search.

Findings from the Literature

Two studies^5,6 were identified from the literature for inclusion in the systematic review (Figure 1). The included studies are summarized in Table 6. A list of excluded studies is presented in Appendix 2.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Description of Studies

Two studies met the inclusion criteria. The FIT1⁵ (N = 76) and FIT2⁶ (N = 74) trials were identically designed double-blind RCTs. Both studies were funded by Rigel Pharmaceuticals Inc. The FIT1 trial was conducted in Australia, Canada (3 sites), Europe, the UK, and the US from July 2014 to April 2016, while the FIT2 trial was conducted in Europe from January 2015 to August 2016. Both trials evaluated the efficacy and safety of fostamatinib compared to placebo among patients with persistent, chronic ITP. Randomization occurred in a 2:1 ratio of fostamatinib to placebo and was stratified by platelet count (< or ≥ 15,000/μL) and a previous splenectomy (yes or no). Before patients were randomized, there was a washout period during which all other ITP therapies were discontinued. The washout period differed according to the specific drug, ranging from 7 days for IVIG to 8 weeks for an alkylating drug. Patients were allowed to continue on corticosteroids (at doses < 20 mg of prednisone equivalent per day), azathioprine, and danazol. In the FIT1 trial, 117 patients were screened and 41 failed screening (3 withdrew consent, 10 failed inclusion criteria, 30 failed exclusion criteria, and 3 failed for other reasons). In the FIT2 trial, 107 patients were screened and 33 failed screening (1 withdrew consent, 15 failed inclusion criteria, 19 failed exclusion criteria, 1 patient was pregnant, and 2 failed for other reasons). In the FIT1 trial, 51 patients were randomized to receive fostamatinib and 25 were randomized to receive placebo, while in the FIT2 trial, 50 patients were randomized to fostamatinib and 24 were randomized to placebo.

Populations

Inclusion and Exclusion Criteria

Both the FIT1 and FIT2 trials enrolled patients 18 years of age and older who had a diagnosis of ITP for at least 3 months and who had received at least 1 previous treatment for ITP. Patients in both trials had to have an average platelet count of less than 30,000/µL from 3 qualifying counts in the previous 3 months, and at least 2 of these counts had to be from the screening period (the 30 days leading up to baseline). Female patients had to be postmenopausal, be surgically sterile, or agree to use an acceptable method of birth control throughout the study. Concurrent treatment for ITP (other than corticosteroids at doses < 20 mg of prednisone equivalent per day, azathioprine, and danazol) was not allowed. Because patients in both trials must have had no known etiology for ITP, those with secondary ITP were excluded. Patients were also excluded if they had a cardiovascular event in the 6 months before randomization, uncontrolled or poorly controlled hypertension (blood pressure ≥ 140/90 mm Hg), surgery in the previous 28 days, any blood products in the 2 weeks before randomization, an infection at screening or baseline, an IBLS score of grade 2, a history of coagulopathy, or an active clinically significant disorder that could affect the pharmacokinetics of fostamatinib.

Baseline Characteristics

In the FIT1 trial, patients in the fostamatinib group were older than the placebo group (a mean age of 57 years versus 53 years), while the proportion of females was lower in the fostamatinib group compared to the placebo group (59% versus 68%, respectively). The patients were predominantly White (86% in fostamatinib group and 84% in placebo) in both FIT1 trial groups. The majority of patients had chronic ITP in both groups, although the proportion was higher in the fostamatinib group compared to the placebo group (94% versus 88%, respectively). Patients in the fostamatinib group had a longer duration of ITP (a mean of 13 years versus 9 years in the placebo group) and had received more previous ITP treatments (a median of 5 compared to 3 in the placebo group). The rate of a previous splenectomy was similar in both groups (39% in fostamatinib and 40% in placebo). Most of the patients had received steroids previously — the proportion was higher in the placebo group compared to the fostamatinib group (100% in placebo versus 90% in fostamatinib). Other common previous treatments included rituximab (51% in fostamatinib versus 44% in placebo), TPO-RAs (51% versus 60%, respectively), and immunoglobulins (65% versus 68%, respectively). The baseline platelet count was higher in the placebo group (a mean of 16,936/µL) compared to the fostamatinib group (a mean of 16,215/µL).

In the FIT2 trial, the baseline age was similar in both groups (a mean of 49 years in the fostamatinib group versus 50 years in the placebo group) while there were more females in the fostamatinib group (62% versus 54% in the placebo group). All patients were White in both groups. Most patients in the FIT2 trial had chronic ITP (94% in fostamatinib group and 96% in placebo group). The duration of ITP was similar between groups (a mean of 12 years in the fostamatinib group and 11 years in the placebo group) while the median number of prior treatments was 3 in both groups. The rate of a previous splenectomy was higher in the placebo group (38%) compared to the fostamatinib group (28%). Most patients had received steroids previously (96% in fostamatinib and 92% in placebo), while other common treatments included TPO-RAs (40% in fostamatinib versus 42% in placebo), immunoglobulins (65% versus 68%, respectively), and azathioprine (34% versus 38%, respectively). The baseline platelet count was higher in the placebo group (a mean of 17,333/µL) compared to fostamatinib (a mean of 15,860/µL).

Patients in the FIT1 trial were older on average than those in the FIT2 trial. Across both trials, most patients were White. The duration of ITP, rate of a previous splenectomy, and use of prior ITP treatments was similar between the FIT1 and FIT2 trials.

Interventions

In both the FIT1 and FIT2 trials, patients randomized to the fostamatinib group received an initial dosage of fostamatinib 100 mg twice daily. Patients in the placebo group took matching placebo. Fostamatinib and matched placebo were taken orally and self-administered. If the platelet count was greater than 50,000/µL, patients remained on the 100 mg twice daily dosage (or matching placebo). If the platelet count was less than 50,000/µL at or after week 4, the dosage of fostamatinib was increased to 150 mg twice daily (or matching placebo) if the patients were tolerating the study drug. The dosage was reduced to fostamatinib 100 mg once daily if patients experienced dose-limiting adverse effects. Patients in both arms were allowed to receive concomitant corticosteroids (equivalent to < 20 mg of prednisone equivalent per day), danazol, or azathioprine. New ITP treatments could not be started, with the exception of rescue therapy. Rescue therapy was permitted for patients with a platelet count below 50,000/µL who were at immediate risk of bleeding or with clinically significant bleeding or wet purpura, or with a platelet count below 50,000/µL and requiring urgent surgery. Rescue therapy could include IVIG, anti-D, or IV methylprednisolone. Beginning at week 12, if the platelet count was less than 50,000/µL (or in the absence of an increase of at least 20,000/µL among patients with a platelet count of less than 15,000/µL at baseline), patients were considered nonresponders and could discontinue participation in the study. Patients could also withdraw from the study if they received rescue therapy after week 10.

Outcomes

A list of efficacy end points identified in the CADTH review protocol that were assessed in the clinical trials included in this review is provided in Table 8. These end points are also summarized in the following section. A detailed discussion and critical appraisal of the outcome measures are provided in Appendix 3.

The primary outcome for both the FIT1 and FIT2 trials was the achievement of a stable platelet response, which was defined as a platelet count of at least 50,000/µL on at least 4 of the last 6 scheduled visits between week 14 and week 24. Secondary efficacy outcomes included achievement of a platelet response (≥ 50,000/µL) at week 12, achievement of a platelet response (≥ 50,000/µL) at week 24, achievement of a platelet count of least 30,000/µL at week 12 among patients with a baseline platelet count of less than 15,000/µL, and achievement of a platelet count of 30,000/µL of greater at week 24 among patients with a baseline platelet count of at less than 15,000/µL. Platelet count was assessed at baseline and every 2 weeks thereafter (± 3 days). Platelet counts were performed by local laboratories affiliated with the clinical sites enrolling patients.

Another secondary outcome was the frequency and severity of bleeding according to the IBLS and the WHO bleeding scale over 24 weeks. This was reported as the mean of IBLS and WHO assessments over 24 weeks per group. Bleeding assessments were performed at baseline and every 2 weeks throughout the study. The IBLS scale assesses bleeding at 9 anatomic sites over the previous week and is rated from 0 (no bleeding) to 2 (marked bleeding). The WHO bleeding scale measures severity of bleeding from 0 (no bleeding) to 4 (debilitating blood loss). Bleeding scale assessments were performed by a physician, a doctor of osteopathic medicine, a physician’s assistant, or a nurse practitioner, and were performed by the same assessor for each patient at all study visits, whenever possible. Another secondary outcome was quality of life, as measured by the SF-36 questionnaire, which is a health survey consisting of 36 questions from 8 HRQoL domains. The SF-36 was completed by patients on their own at baseline, week 4, week 12, and week 24. For exploratory post hoc analyses, the investigators measured the onset of platelet effect in responders, use of rescue therapy, and bleeding-related serious adverse effects. The investigators also conducted exploratory analyses of subgroups according to age (< 57 years or ≥ 57 years), sex, prior experience with TPO-RA, a prior splenectomy, and baseline platelet counts (< 15,000/µL or ≥ 15,000/µL).

The clinical study reports for the FIT 1 and FIT2 trials state that all efficacy and safety measures were standard (i.e., they were widely used and generally recognized as reliable, accurate, and relevant). Contract research organizations were responsible for study conduct, data management, and statistical analysis. An electronic case report form was used to record data, which was completed by study staff. A safety contract research organization set up the safety database and was responsible for data entry, coding, and preparation of safety reports.

Treatment-emergent adverse events were either new or detectable exacerbations of pre-existing conditions. The AE reporting period began with the first dose of study and ended with the final study visit. Serious adverse events were any untoward medical occurrences that resulted in death, were life-threatening, required inpatient hospitalization, or resulted in significant disability or incapacity. Investigators assessed the occurrence of AEs and SAEs at all evaluation time points during the study. When AE or SAEs were volunteered by patients, discovered by staff during questions, or detected by physical exams or laboratory tests, the AEs were recorded in a case report form and coded using the Medical Dictionary for Regulatory Activities.

Statistical Analysis

The primary efficacy outcome for the FIT1 and FIT2 trials was a stable platelet response as defined previously. Patients who discontinued treatment before week 24 because of a lack of efficacy or an AE, or who received rescue treatment after 10 weeks, were considered nonresponders. The proportion of patients achieving a stable platelet response in each treatment group was calculated for the ITT population. The authors reported the RD and 95% CI around the RD based on the normal approximation. The null hypothesis was that there was no difference in the proportion of responders for the fostamatinib group compared to placebo. The investigators used a 2-sided Fisher exact test to test the null hypothesis, with a significance level of 0.05. The sample size was calculated to provide a 90% power for the primary efficacy outcome, using a 2:1 ratio for fostamatinib: placebo, and assuming the proportion of responders in the fostamatinib group would be 0.4 and the proportion of responders in the placebo group would be 0.05. The required sample size was 75 patients. The last observation carried forward (LOCF) method was used to impute missing data for the primary outcome. The LOCF was used to impute missing platelet counts for patients who withdrew from the study early (due to reasons other than lack of efficacy, use of rescue therapy, or an AE; patients withdrawing for those reasons were deemed nonresponders). In the FIT2 trial, the investigators also used multiple imputation as a sensitivity analysis of the primary outcome when there were missing platelet count data. This analysis was stated to be pre-planned in the Clinical Study Report for FIT2; however, no information on the methods for this analysis were identified in the Clinical Study Report nor protocol. In the results section of the FIT2 Clinical Study Report, the investigators stated that missing platelet counts were imputed using the SAS MINANALYZE procedure, and reported proportions were the average of 10,000 iterations.

Secondary efficacy outcomes involving platelet counts were analyzed the same way as the primary outcome (including data imputation using LOCF). For the IBLS and WHO bleeding scale data, the mean of the assessment scores was calculated across the 24-week treatment period. The investigators used a 2-sided, 2-sample t-test to evaluate the difference in means. Descriptive statistics were used to describe rescue therapy use and bleeding-related SAEs; however, no statistical test was used to evaluate differences between treatment groups. Exploratory subgroup analyses for the primary outcome involved calculating the proportion achieving the primary efficacy end point in each treatment group within the subgroup, calculating an RD, and reporting a 95% CI, around the RD. No statistical test was performed to evaluate differences between treatment arms in different subgroups.

For AEs, the investigators reported the number and proportion of patients with at least 1 AE during the double-blind treatment period.

Analysis Populations

In the FIT1 and FIT2 trials, the ITT population included all randomized patients. The ITT population was used for analysis of all efficacy analyses and patients were analyzed according to their randomized treatment assignment. The per-protocol population was all patients with no major protocol violations (those not receiving any study treatment, not receiving correct study treatment, or failing to meet eligibility criteria). All outcomes analyzed in the per-protocol population were according to patients’ randomized treatment assignment. The safety population was all randomized patients who received any amount of study drug, and patients were analyzed according to the treatment they received.

Results

Patient Disposition

In the FIT1 trial, 117 patients were screened, and 76 patients (65%) were randomized — 51 to fostamatinib and 25 to placebo. A total of 39 patients (76%) in the fostamatinib group discontinued the study early, while 24 patients (96%) in the placebo group discontinued early. In the FIT2 trial, 107 patients were screened and 74 (69%) were randomized — 50 to fostamatinib and 24 to placebo. A total of 37 patients (74%) in the fostamatinib group discontinued from the study early, while 22 patients (92%) in placebo group discontinued early. Reasons for discontinuation from the FIT1 and FIT2 trials are listed in Table 9. In the FIT1 trial, 55% of patients in the fostamatinib group discontinued early due to a lack of response compared to 88% in the placebo group. In the FIT2 trial, 66% of patients in the fostamatinib group discontinued early due to a lack of response compared to 79% in the placebo group.

Exposure to Study Treatments

In the FIT1 trial, the mean duration of exposure was 99 days (SD = 46) for fostamatinib and 94 days (SD = 30) for placebo. The mean overall compliance (actual number of tablets taken divided by total number expected to be taken) was 93% in the fostamatinib group and 96% in the placebo group. In the fostamatinib group, 2% of patients received concomitant danazol, 4% received concomitant platelets, 37% received concomitant steroids, and 29% received concomitant immunoglobulins. In the placebo group, 56% of patients received concomitant steroids, 4% of patients received concomitant platelets, and 28% of patients received concomitant immunoglobulins. Use of rescue therapy is reported as an efficacy outcome.

In the FIT2 trial, the mean duration of exposure was 112 days (SD = 41) in the fostamatinib group and 88 days (SD = 36) in the placebo group. The mean overall compliance was 99% in the fostamatinib group and 98% in the placebo group. In the fostamatinib group, 2% of patients received concomitant danazol, 44% received concomitant steroids, and 12% received concomitant immunoglobulins. In the placebo group, 4% of patients took concomitant platelets, 63% of patients received concomitant steroids, and 25% of patients received concomitant immunoglobulins. Use of rescue therapy is reported as an efficacy outcome.

Efficacy

Only those efficacy outcomes and analyses of subgroups identified in the review protocol are reported here; Appendix 3 provides detailed efficacy data. Pooled outcome data for the FIT1 and FIT2 trial were reported in a single publication.²³ These data are also presented in Appendix 3. The following outcomes identified as important in the review protocol were not measured in the FIT1 trial nor the FIT2 trial: emergency room visits, hospitalizations, treatment-free remission, and symptoms.

Stable Platelet Response

In the FIT1 trial, 18% of patients in the fostamatinib group experienced a stable platelet response compared to 0% in the placebo group (RD = 18%; 95% CI, 7.2 to 28; P = 0.026).

In the FIT2 trial, 18% of patients in the fostamatinib group experienced a stable platelet response compared to 4% in the placebo group (RD = 14%; 95% CI, 0.5 to 27; P = 0.15).

Platelet Count of 50,000/µL or Greater

In the FIT1 trial, 22% of patients in the fostamatinib group achieved a platelet count of at least 50,000/µL at week 12 compared to 0% in the placebo group (RD = 22%; 95% CI, 10 to 33; P = 0.013). At week 24, 16% of patients in the fostamatinib group achieved a platelet count of at least 50,000/µL compared to 0% in the placebo group (RD = 16%; 95% CI, 5.7 to 26; P = 0.047).

In the FIT2 trial, 24% of patients in the fostamatinib group achieved a platelet count of at least 50,000/µL at week 12 compared to 13% in the placebo group (RD = 12%; 95% CI, −6.3 to 29; P = 0.36). At week 24, 16% of patients in the fostamatinib group achieved a platelet count of at least 50,000/µL compared to 4% in the placebo group (RD = 12%; 95% CI, −1.1 to 25; P = 0.26).

Platelet Count of 30,000/µL or Greater and 20,000/µL or Greater Above Baseline in Patients With a Low Baseline Platelet Count (Less Than 15,000/µL)

In the FIT1 trial, 16% of patients with a low baseline platelet count in the fostamatinib group achieved a platelet count of at least 30,000/µL and at least 20,000/µL above baseline at week 12 compared to 0% in the placebo group (RD = 16%; 95% CI, 1.6 to 30; P = 0.28). At week 24, 16% of patients with a low baseline platelet count in the fostamatinib group achieved a platelet count of 30,000/µL or greater and 20,000/µL or greater above baseline compared to 0% in the placebo group (RD = 16%; 95% CI, 1.6 to 30; P  = 0.28).

In the FIT2 trial, 27% of patients with a low baseline platelet count in the fostamatinib group achieved a platelet count of 30,000/µL or greater and 20,000/μL or greater above baseline at week 12 compared to 11% in the placebo group (RD = 16%; 95% CI, −12 to 44; P = 0.64). At week 24, 14% of patients with a low baseline platelet count in the fostamatinib group achieved a platelet count of 30,000/µL or greater and 20,000/μL or greater above baseline compared to 0% in the placebo group (RD = 14%; 95% CI, −0.7 to 28; P = 0.54).

Bleeding Assessments

In the FIT1 trial, the mean IBLS score across 24 weeks was 0.13 (SD = 0.12) in the fostamatinib group and 0.14 (SD = 0.10) in the placebo group (difference in means = −0.01; 95% CI, −0.1 to 0.0; P = 0.66). The mean WHO bleeding scale score across 24 weeks in the fostamatinib group was 0.61 (SD = 0.66) compared to 0.46 (SD = 0.56) in the placebo group (difference in means = 0.15; 95% CI, −0.2 to 0.5; P = 0.34).

In the FIT2 trial, the mean IBLS score across 24 weeks was 0.04 (SD = 0.08) in the fostamatinib group and 0.06 (SD = 0.07) in the placebo group (difference in means = −0.01; 95% CI, −0.05 to 0.02; P = 0.49). The mean WHO bleeding scale score across 24 weeks in the fostamatinib group was 0.26 (SD = 0.38) compared to 0.38 (SD = 0.47) in the placebo group (difference in means = −0.12; 95%−CI, −0.32 to 0.09; P = 0.25).

Use of Rescue Therapy

In the FIT1 trial, 31% of patients in the fostamatinib group required rescue therapy before week 10 compared to 44% of patients in the placebo group. After week 10, 14% of patients in the fostamatinib group required rescue therapy compared to 28% in the placebo group.

In the FIT2 trial, 18% of patients in the fostamatinib group required rescue therapy before week 10 compared to 29% of patients in the placebo group. After week 10, 2% of patients in the fostamatinib group required rescue therapy compared to 21% in the placebo group.

Bleeding-Related Serious Adverse Events

In the FIT1 trial, ||| of patients in the fostamatinib group experienced a bleeding-related SAE compared to ||| in the placebo group. In the FIT2 trial, ||| of patients in the fostamatinib group experienced a bleeding-related SAE compared to ||| in the placebo group.

Short Form (36) Health Survey

In the FIT1 trial, there were no differences in SF-36 scores between the fostamatinib and placebo groups at any time point. At week 24, there was ||||||||| providing SF-36 data in the placebo group and |||| patients in the fostamatinib group.

In the FIT2 trial, there were no differences in SF-36 scores between the fostamatinib and placebo groups at week 12 or week 24; however, at week 4 the mean change from baseline in the fostamatinib group for bodily pain was |||| compared to |||| in the placebo group (difference in mean change from baseline = ||||||||||||). At week 4 the mean change from baseline for general health (difference in mean change from baseline = ||||||||||||) and physical health (difference in mean change from baseline = |||||||||||) was also greater in the fostamatinib group compared to placebo. At week 24 in the FIT2 trial, there were |||||||||||| providing SF-36 data in the placebo group and |||| patients in the fostamatinib group.

Time to Response Among Responders

In the FIT1 trial, the mean time to a platelet count of 50,000/µL or greater among responders was 39 days (range = 15 to 73 days; median not reported). In the FIT2 trial, the mean time to a platelet count of 50,000/µL or greater among responders was 22 days (median = 15 days; range = 12 to 56 days).

Subgroup Analysis

Both the FIT1 and FIT2 trials conducted subgroup analyses for the primary efficacy end point. Results are presented in Table 11.

In the FIT1 trial, among patients with prior experience with TPO-RA treatment, 15.4% of patients in the fostamatinib group experienced a stable platelet response compared to 0% in the placebo group (RD = 15.4%; 95% CI, 1.5 to 29.3). Among patients without prior experience with TPO-RA treatment, 20% of patients in the fostamatinib group experienced a stable platelet response compared to 0% in the placebo group (RD = 20%; 95% CI, 4.3 to 35.7). In the FIT2 trial, among patients with prior experience with TPO-RA treatment, 15% of patients in the fostamatinib group experienced a stable platelet response compared to 0% in the placebo group (RD = 15%; 95% CI, −0.6 to 30.6). Among patients without prior experience with TPO-RA treatment, 20% of patients in the fostamatinib group experienced a stable platelet response compared to 7% in the placebo group (RD = 12.9%; 95% CI, −6.8 to 32.5).

In the FIT1 trial, among patients with a prior splenectomy, 15% of patients in the fostamatinib group experienced a stable platelet response compared to 0% in the placebo group (RD = 15%; 95% CI, −0.6 to 30.6). Among patients without a prior splenectomy, 19.4% of patients in the fostamatinib group experienced a stable platelet response compared to 0% in the placebo group (RD = 19.4%; 95% CI, 5.4 to 33.3). In the FIT2 trial, among patients with a prior splenectomy, 21.4% of patients in the fostamatinib group experienced a stable platelet response compared to 0% in the placebo group (RD = 21.4%; 95% CI, −0.1 to 42.9). Among patients without a prior splenectomy, 16.7% of patients in the fostamatinib group experienced a stable platelet response compared to 6.7% in the placebo group (RD = 10%; 95% CI, −7.5 to 27.5).

Table 11: Subgroup Analyses for Primary Efficacy Outcome (Intention-to-Treat Population)

CI = confidence interval; NR = not reported; TPO-RA = thrombopoietin receptor agonist.

^aIntention-to-treat population, no statistical test performed.

Source: FIT1 Clinical Study Report⁵ and FIT2 Clinical Study Report.⁶

Subgroup analyses of the stable platelet response rate (primary efficacy end point) based on data pooled from the 2 studies support the consistency of the efficacy of fostamatinib across subpopulations based on disease-related characteristics (baseline platelet count, prior splenectomy, previous use of a TPO-RA, and previous use of rituximab) and demographics (age or sex). Patients who had ITP for less than 3 years did not respond to fostamatinib as well as those who had ITP for more than 3 years; however, these results could have been affected by the small number of patients who had ITP for less than 3 years. Results appear in Figure 2.

Figure 2: Subgroup Analyses for the Primary Efficacy Outcome of Stable Platelet Response at Week 24 Pooled From FIT1 and FIT2

CI = confidence interval; ITP = immune thrombocytopenia; TPO = thrombopoietin.

Source: Tavalisse Drug Reimbursement Review submission.²⁴

Harms

Only those harms identified in the review protocol are reported. Table 12 provides detailed harms data.

Adverse Events

In the FIT1 trial, 96% of patients in the fostamatinib group experienced an AE compared to 76% of patients in the placebo group. In the fostamatinib group, the most common AEs were diarrhea (41%), nausea (29%), increased ALT (18%), increased AST (16%), headache (14%), dizziness (18%), epistaxis (18%), fatigue (12%), and hypertension (26%). The most common AEs in the placebo group were diarrhea (16%), headache (24%), dizziness (16%), epistaxis (16%), and dyspnea (12%).

In the FIT2 trial, 71% of patients in the fostamatinib group experienced an AE compared to 78% in the placebo group. The most common AEs in the fostamatinib group were diarrhea (18%), epistaxis (12%), and hypertension (14%). The most common AEs in the placebo group were diarrhea (13%), nausea (13%), headache (13%), hypertension (13%), and thrombocytopenia (13%).

Serious Adverse Events

In the FIT1 trial, 16% of patients in the fostamatinib group had at least 1 SAE compared to 20% in the placebo group. In the fostamatinib group, 1 patient (2%) experienced febrile neutropenia, 1 patient experienced immune thrombocytopenic purpura, 1 patient a retinal tear, 1 patient diarrhea, 1 patient pneumonia, 1 patient syncope, 1 patient vaginal hemorrhage, and 1 patient epistaxis. In the placebo group, 1 patient (4%) experienced anemia, 1 patient cardiac congestive failure, 1 patient gastrointestinal hemorrhage, 1 patient sepsis, 1 patient menorrhagia, 1 patient chronic obstructive pulmonary disease, and 1 patient epistaxis.

In the FIT2 trial, 10% of patients in the fostamatinib group had at least 1 SAE compared to 26% in the placebo group. In the fostamatinib group, 1 patient (2%) had epistaxis, 1 patient had bronchitis, 1 had a contusion, 1 a platelet count decrease, 1 plasma cell myeloma, 1 a transient ischemic attack, and 1 a hypertensive crisis. In the placebo group, 3 patients (13%) had thrombocytopenia, 1 menorrhagia, 1 a muscle rupture, 1 an infection, and 1 petechiae.

Withdrawals Due to Adverse Events

In the FIT1 trial, 16% of patients in the fostamatinib group withdrew due to any AE compared to 8% in the placebo group. In the fostamatinib group, 1 patient (2%) withdrew due to abdominal pain, 1 due to diarrhea, 1 due to neutropenia, 1 due to thrombocytopenia, 1 due to increased ALT, 1 due to chest pain, 1 due to pneumonia, and 1 due to syncope. In the placebo group, 1 patient (4%) withdrew due to abdominal discomfort and 1 due to epistaxis.

In the FIT2 trial, 4% of patients in the fostamatinib group withdrew due to any AE compared to 9% in the placebo group. In the fostamatinib group, 1 patient (2%) withdrew due to plasma cell myeloma and 1 due to headache. In the placebo group, 1 patient (4%) withdrew due to diarrhea and 1 due to hypertension.

Mortality

In the FIT1 trial, 1 patient died in the placebo group due to sepsis. In the FIT2 trial, 1 patient died in the fostamatinib group due to plasma cell myeloma.

Notable Harms

In the FIT1 trial, ||||| of patients in the fostamatinib group experienced an infection compared to ||| in the placebo group. In the FIT2 trial, ||||| of patients in the fostamatinib group and ||| of patients in the placebo group experienced an infection. In both the FIT1 and FIT2 trials, |||| of patients in the fostamatinib group experienced neutropenia compared to |||| in the placebo group. In the FIT1 trial, ||||| of patients in the fostamatinib group experienced elevated liver transaminase levels compared to |||| in the placebo group. In the FIT2 trial, |||| of patients in the fostamatinib group experienced elevated liver transaminase levels compared to |||| in the placebo group.

Table 12: Summary of Harms

AE = adverse event; ALT = alanine transaminase; AST = aspartate transaminase; COPD = chronic obstructive pulmonary disease; NA = not applicable; TIA = transient ischemic attack; URTI = upper respiratory tract infection; UTI = urinary tract infection.

^aFrequency greater than 5%.

^bFrequency greater than 2%.

Source: FIT1 Clinical Study Report,⁵ and FIT2 Clinical Study Report.⁶

Critical Appraisal

Internal Validity

Both the FIT1 and FIT2 trials were double-blind and, according to the clinical study reports, the unblinding procedure was not used during the studies; however, unblinding for patients and clinicians may have been possible as a platelet response in patients receiving placebo would not likely be observed. There was high adherence to the study treatments in both groups. Overall, the FIT1 and FIT2 trials were at low risk of bias due to protocol deviations from intended interventions. In the FIT1 and FIT2 trials, outcome data were available for nearly all participants randomized for the primary outcome; bias from missing outcome data for the primary outcome was therefore unlikely. However, there were missing outcome data for the HRQoL outcome. The measurement of outcomes in the FIT1 and FIT2 trials was the same in both groups. Outcome assessors were not aware of the intervention received by patients and, even if they were aware, it would be unlikely to influence assessment of platelet counts. The FIT1 and FIT2 trials were therefore at low risk of bias for measurement of outcomes. In the FIT1 trial, data were reported according to a pre-specified analysis plan, and post hoc analyses were acknowledged as such and described in the methods, with the exception of subgroup analyses that were not described in the methods or protocol. In the FIT2 trial, data were generally reported according to a pre-specified plan; however, a sensitivity analysis involving multiple imputation for the primary outcome was reported in the results. This analysis was claimed to be pre-specified, although there was no description in the methods section of the clinical study report or protocol. Similar to the FIT2 trial, subgroup analyses were not described in the methods or protocol, creating some concerns in the FIT1 and FIT2 trials related to selective outcome reporting.

In both the FIT1 and FIT2 trials, treatment groups were generally balanced in terms of baseline characteristics such as platelet count, rate of previous splenectomy, and age, although some imbalances existed. In the FIT1 trial, there was a higher proportion of females in the placebo group (68%) compared to the fostamatinib group (59%), while in the FIT2 trial the proportion of females was higher in the fostamatinib group (62% versus 54% in placebo). In the FIT1 trial, the fostamatinib group had a slightly longer duration of ITP on average (a mean of 13 years versus 9 years in the placebo group) and had used a greater number of previous ITP therapies (a median of 5 versus 3 in placebo group) compared to placebo, although this was likely not of major prognostic importance as both groups had longstanding ITP and were heavily pre-treated. There were also possible imbalances in the rates of specific treatments used in the FIT1 trial — for example, the rate of prior steroid use was higher in the placebo group versus fostamatinib (100% versus 90%, respectively), along with the rate of prior TPO-RA use (60% in placebo versus 51% in fostamatinib). In the FIT2 trial, the rate of prior splenectomy was higher in the placebo group (38%) compared to the fostamatinib group (28%). Subgroup analyses did not reveal differences in the response rate based on prior use of TPO-RAs or a splenectomy. However, these analyses were likely underpowered, and it is possible that baseline imbalances in prior treatments used may have introduced bias. Further, in both the FIT1 and FIT2 trials, the rate of concomitant steroid use was higher in the placebo group compared to the fostamatinib group, and rescue therapy use was higher in the placebo group as well. This may have introduced bias against the fostamatinib treatment group. The rate of study discontinuation was high in both the FIT1 and FIT2 trials and was imbalanced between study groups in both trials. The most common reason for discontinuation was a lack of treatment response after week 12 (e.g., 55% in the fostamatinib group and 88% in the placebo group of the FIT1 trial). In both groups, these patients were deemed to be nonresponders. However, for the SF-36 outcome, the high study-discontinuation rate meant limited data were available at week 24 (e.g., 1 patient in the placebo group at week 24 in the FIT1 trials, and 2 patients in the placebo group at week 24 in the FIT2 trial). Differences in the SF-36 outcome favoured fostamatinib at week 4; however, the differences were not likely clinically meaningful as they occurred so early in the trial, and it was not possible to draw any meaningful conclusions from the SF-36 data at week 24 due the limited amount of data from study discontinuations.

Imputation of the LOCF was used in both the FIT1 and FIT2 trials for missing platelet count data. Platelet count data were imputed for 3 patients in the fostamatinib group (all deemed to be nonresponders) in the FIT1 trial. In the FIT2 trial, data were imputed for 2 patients (1 responder and 1 nonresponder) in the fostamatinib group and 1 patient in the placebo group (deemed a nonresponder). The 1 patient in the fostamatinib group deemed a responder had a baseline platelet count of 5,000/µL. The count rose to 84,000/µL at week 12, 99,000/µL at week 14, and 101,000/µL at week 16, after which point no platelet counts were available. The effect of using the LOCF for imputation in the FIT and FIT2 trials therefore did not appear to have a major impact on the outcomes.

Subgroup analyses were not pre-specified or adjusted for multiplicity. Given the small number of patients in each subgroup and low event rates, there was likely insufficient power to detect any differences between treatment groups in these subgroups. This is reflected by wide CIs in the RD. The small number of patients and low event rates for certain outcomes (bleeding-related SAEs and use of rescue therapy) make it challenging to draw conclusions surrounding any difference between treatment arms for these secondary end points. As mentioned previously, the high study-discontinuation rate meant there was a large amount of missing data for the SF-36 outcome, making it impossible to draw conclusions about differences between treatment groups for this outcome. Because neither the FIT1 trial nor the FIT2 trial were powered for secondary end points and there was no adjustment for multiplicity for secondary end points, these outcomes should be interpreted with caution.

External Validity

The clinical experts consulted by CADTH indicated that the population of the FIT1 and FIT2 trials are broadly comparable to the population of patients with ITP in Canada, and the results of these trials are likely generalizable in Canada. The long duration of ITP and multiple previous treatments among patients in the FIT2 trial mirrors what is commonly seen in patients with ITP in clinical practice in Canada. However, the clinical experts did note some generalizability concerns with the FIT1 and FIT2 trials. First, because the FIT1 and FIT2 participants were predominantly White, the results from these trials may not be generalizable to other racial groups commonly seen at some centres in Canada (although the experts noted that the role of race and/or ethnicity in treatment response was not clear). The experts also noted that, because patients with secondary ITP were excluded from the FIT1 and FIT2 trials, the trial findings may not be generalizable to those with secondary ITP. Further, the specific types of previous treatments used in the FIT1 and FIT2 trials differ from those commonly seen at a similar point in ITP treatment in Canada. The clinical experts pointed out that, based on the duration of ITP for patients in the FIT1 and FIT2 trials, a higher portion of chronic ITP patients in Canada would have had a prior splenectomy. Moreover, the extent of previous use of rituximab in the FIT1 trial is higher than what would be seen in Canada at a similar stage of treatment. Nevertheless, the high median number of previous therapies used over the course of ITP reflects what would be seen in Canadian practice. The clinical experts also noted that the co-interventions (i.e., concomitant ITP medication use) during the FIT1 and FIT2 trials are also reflective of real-world practice in Canada, as steroids would commonly be continued while a patient is initiated on a new long-term treatment strategy such as fostamatinib.

In terms of outcome assessment in the FIT1 and FIT2 trials, the clinical experts noted that, while bleeding outcomes are likely the most important in practice, treatment response is most commonly assessed by measuring platelet counts. The experts also stated that having serial platelet measurements at weeks 14 to 24 is reflective of how patients would be followed in clinical practice. The approach used to assess platelets in the FIT1 and FIT2 trials is therefore generalizable. However, the experts found the definition for response to therapy in the FIT1 and FIT2 trials to be overly strict, stating that achieving a platelet count exceeding 30,000/µL would be considered a response in practice, particularly if a patient is asymptomatic. As a result, the primary outcome definition used in the FIT1 and FIT2 trials may not be reflective of how response to therapy is assessed in routine practice.

The FIT1 and FIT2 trials provided limited data on clinically important outcomes such as quality of life, rescue therapy, and bleeding events. The clinical experts do not use the IBLS and WHO bleeding scales in practice, and the relevance of the bleeding outcome scales used in the FIT1 and FIT2 trials is unclear. Further, the event rates for the post hoc bleeding-related SAE outcome made it challenging for the clinical experts to comment on the relevance or meaningfulness of these findings. The clinical experts found that the lower rate of rescue therapy use among patients treated with fostamatinib compared to placebo in the FIT1 and FIT2 trials could be meaningful, but they also noted the relatively low event rates.

Another challenge with both the FIT1 and FIT2 trials is that the comparator is placebo. In chronic ITP, if a platelet count is below 20,000/µL, as it was at baseline for patients in both trials, the clinical experts (and clinical practice guidelines) indicate that treatment would be warranted. Placebo therefore may not be an appropriate comparator for fostamatinib. Indeed, the FIT1 and FIT2 trials do not address the comparative efficacy of fostamatinib against other second- or third-line ITP treatments. However, the clinical experts suggested that the response rate seen in the FIT1 trial and FIT2 with fostamatinib among heavily pre-treated patients is meaningful for those who have taken several prior therapies and have a long duration of ITP.

Indirect Evidence

Objectives and Methods for the Summary of Indirect Evidence

As there was no direct evidence comparing fostamatinib to other active therapies for the treatment of chronic ITP in adult patients who have had an insufficient response to other treatments, a review of indirect evidence was undertaken. In addition to reviewing the sponsor’s submission, CADTH conducted a literature search to identify potentially relevant ITCs in patients with ITP. A focused literature search for ITCs and NMAs dealing with ITP was run in MEDLINE All (1946–) on August 20, 2021. No limits were applied to the search. Titles, abstracts, and full-text articles were screened for inclusion by 1 reviewer based on the population, intervention, comparator, and outcome criteria outlined in Table 5.

One sponsor-submitted ITC was included in this review.⁸ One relevant ITC by Wojciechowski et al. was identified in the literature search.⁷

Description of Indirect Comparisons

Both the sponsor-submitted ITC and the Wojciechowski study included a systematic review of the literature and an ITC that compared the current pharmaceutical treatments with each other for chronic ITP. In the sponsor-submitted ITC, fostamatinib was compared to rituximab. In the Wojciechowski study, fostamatinib was compared to 3 TPO-RAs: avatrombopag, eltrombopag, and romiplostim.

Methods of the Sponsor-Submitted Indirect Treatment Comparison

Objectives

The objective of the sponsor-submitted report for patients with chronic ITP was to conduct a systematic review and, if possible, an ITC to evaluate the relative efficacy and safety of fostamatinib versus other treatments currently available in Canada and Europe for this population. Chronic ITP was defined as ITP with a duration of at least 12 months.

Study Selection Methods

A literature systematic review was performed in 2019 and updated in 2021 to identify all relevant clinical evidence to inform the NMA of fostamatinib in patients with chronic ITP. Multiple databases were searched. The inclusion and exclusion criteria used for study selection to inform the NMA are presented in Table 13. Study selection and data extraction were conducted by 2 reviewers independently. The quality of the included studies was not assessed.

Methods of the Indirect Treatment Comparison by Wojciechowski et al.

Objectives

The objective of this report was to conduct an ITC to evaluate the relative efficacy and safety of avatrombopag, relative to eltrombopag, romiplostim, and fostamatinib, for patients with chronic ITP who were not responding adequately to corticosteroids. Chronic ITP was defined as ITP with a duration of at least 12 months.

Study Selection Methods

A systematic literature search was conducted to identify RCTs and observational studies involving adult patients with chronic ITP. Multiple databases were searched to identify clinical trials that evaluated the efficacy and safety of drug therapies for chronic ITP. By definition, chronic ITP is required to have a duration of at least 12 months; however, some included trials may have been designed and conducted before the current definition of chronic ITP was developed; and patients with a shorter disease duration, e.g., at least 6 months, may have been recruited. Such studies were deemed eligible for this analysis if all other inclusion criteria for the NMA were met, and the average duration of the disease was at least 12 months.

Inclusion and exclusion criteria for the clinical studies for the 2 systematic reviews are presented in Table 13.

Table 13: Study Selection Criteria and Methods for Indirect Treatment Comparisons

AE = adverse event; CRD = Centre for Reviews and Dissemination; EMA = European Medicines Agency; HRQoL = health-related quality of life; ITC = indirect treatment comparison; ITP = immune thrombocytopenia; PRISMA = Preferred Reporting Items for Systematic Review and Meta-Analyses; RCT = randomized controlled trial; SAE = serious adverse event.

Source: Sponsor-submitted indirect treatment comparison⁸ and Wojciechowski et al. (2021).⁷

Analysis Methods of the Sponsor-Submitted Indirect Treatment Comparison

In the feasibility assessment of this study, the following patient baseline characteristics were considered potential treatment-effect modifiers in patients receiving fostamatinib or rituximab:

• age

• gender

• ethnicity

• time since ITP diagnosis

• prior TPO-RAs treatment

• prior rituximab treatment

• baseline platelet count

• concomitant medications

• prior splenectomy.

The clinical experts consulted by the sponsor indicated that none of these pre-specified patient characteristics could be considered reliable treatment-effect modifiers in the study population due to a lack of evidence.

In the analyses, the probabilities of platelet response were modelled using a logit link function. A random-effects model within the Bayesian framework was selected. In this model, prior distributions for the parameters were specified. Where sufficient sample data were available, conventional reference prior distributions were used:

Trial-specific baseline:

Treatment effects relative to reference treatment:

Between-study SD of treatment effects:

In the case of insufficient sample data in a network, the prior for τ² proposed by Ren et al.²⁵ was used, which was a truncated prior (a lognormal [−2.56, 1.74²]). The truncation was based on the judgment that the hazard ratio divided by the OR in 1 study would be no more than 10 times greater than in another. In the case of 0 events, a continuity correction was applied by adding 1 to the denominator and 0.5 to the numerator.

For all outcomes, a burn-in of 70,000 iterations of a Markov chain was used, with a further 50,000 iterations retained to estimate parameters using 1 chain and thinning every 5 iterations.

The model fit was examined by comparing the total residual deviance to the total number of data points included in the analysis and deviance information criterion. The results showed that the model fit the data well, with a total residual deviance of 12.66 being close to the number of data points (13) included in the analysis. Overall platelet response was assessed in this ITC and reported as the OR, which represents the relative likelihood of achieving a platelet response when comparing 1 received therapy against another. The results were presented using the posterior median treatment effects (95% Cr). The estimated between-study SD, τ, for each analysis was also presented. Values below 0.05 were considered to indicate low heterogeneity. Values between 0.05 and 0.5 were considered to indicate moderate heterogeneity. Values between 0.5 and 1.0 were considered to indicate high heterogeneity. Values above 1.0 were considered to indicate extremely high heterogeneity.

Three analyses were performed in this ITC:

• Analysis 1 used a definition of overall platelet response from each study publication and included the results from all included publications. In this analysis, the outcome of interest in the FIT1 and FIT2 trials was more than 1 platelet count of 50,000/µL or greater during weeks 0 to 12, while the outcome of interest was a platelet count exceeding 30,000/µL in other included trials.

• Analysis 2 used alternative definitions of overall platelet response for the FIT1 and FIT2 trials, defined as a platelet count greater than 30,000/µL by week 4 of treatment, and including the results from all 4 rituximab publications. In this analysis, the outcome of interest was a platelet count of exceeding 30,000/µL at the study end.

• Analysis 3 involved a comparison to the Ghanima study (2015) only, using the same definition for outcomes as analysis 2 (a platelet count greater than 30,000/µL by week 4 of treatment). In this analysis, the outcome of interest was a platelet count of greater than 30,000/µL at the study end.

All analyses in this study were conducted using the software package WinBUGS and R.

Indirect Treatment Comparison Analysis Methods in Wojciechowski et al.

The NMA was conducted within a Bayesian framework, and a Markov chain Monte Carlo method was implemented in WinBUGS with vague prior distributions for model parameters. Noninformative prior distributions were used for the model’s nuisance parameters, treatment-effect parameters (normal distributions with a mean of 0 and a variance of 10⁴), and heterogeneity parameters (uniform distribution between 0 and 5 for between-trial SD) for the Bayesian analysis.

Fixed-effects and random-effects models were fitted to the data, with model fit based on the deviance information criterion. In this study, the fixed-effects model was preferred when fewer estimable parameters were evaluated. Three chains were run for each analysis, with either 25,000 or 50,000 burn-in iterations for the fixed-effects and random-effects models, respectively, followed by 25,000 iterations. Trace plots were generated to assess convergence. The consistency of the results within closed loops of the network was tested using a modified Bucher approach.

Outcomes were presented as ORs or incidence rate ratios with corresponding 95% CrIs. The significant and imbalanced discontinuation introduced a high risk of bias. To adjust for this imbalance due to discontinuation, the NMA was conducted using estimated incidence rate ratios for the need for rescue therapy and the incidence of any bleeding events, WHO grade 2 to 4 bleeding events, and any AEs.

Table 14: Indirect Treatment Comparison Analysis Methods

DIC = deviance information criterion; ITC = indirect treatment comparison; NA = not available; NR = not reported.

Source: Sponsor-submitted indirect treatment comparison⁸ and Wojciechowski et al.⁷

Results of Indirect Treatment Comparisons

Summary of Included Studies

Sponsor-Submitted Indirect Treatment Comparison

From the literature systematic review, 3,233 studies were identified, and among them, a total of 31 studies were evaluated in a feasibility assessment to determine if it was possible to conduct an ITC in the study population. Six RCTs were included and contributed evidence on overall platelet response for the ITC. Authors of the ITC indicated fostamatinib was included in the treatment paradigms to address patients with the greatest unmet need. Patients who were successfully treated with a splenectomy or TPO-RAs were unlikely to require additional treatment with a new intervention. To address the patients with the greatest unmet need, fostamatinib focused on those patients who did not have access to long-term effective therapy options, such as those who received short courses of rituximab and those on a watch-and-rescue regimen.

In the trials included in ITC, the number of enrolled patients ranged from 57 to 138. The trial duration ranged from 4 weeks to 78 weeks. Three doses of rituximab were evaluated: 2 or 4 once-weekly 375 mg/m² doses, 2 once-weekly 750 mg/m² doses, or 4 once-weekly 100 mg/m² doses (Figure 3). The definition of platelet response varied across the included trials.

Figure 3 presents the network of evidence for overall platelet response in patients with chronic or persistent ITP in the sponsor-submitted ITC.

Figure 3: Network of Evidence for Platelet Response

Source: Sponsor-submitted indirect treatment comparison.⁸

The heterogeneity of the included RCTs was assessed to determine whether an NMA would have been possible. When comparing patient baseline characteristics, the average age was considered homogenous across all studies and ranged for most trials between 40 and 57 years. The average duration of ITP was heterogenous in the included studies. The baseline platelet counts were similar across trials (Table 16).

Wojciechowski Study

A total of 1,822 publications were identified from the literature search, and among them, 64 studies were evaluated in a feasibility assessment to determine if it was possible to conduct an ITC in the study population. Seven phase III, double-blind RCTs were included and contributed evidence on various clinical outcomes to the ITC (Table 17): 2 for avatrombopag, 1 for eltrombopag, 2 for romiplostim, and 2 for fostamatinib. Six of them were placebo-controlled trials, and 1 was designed to compared avatrombopag versus eltrombopag.

Figure 4 presents the network of evidence for overall platelet response in patients with chronic or persistent ITP in the Wojciechowski ITC.

Figure 4: Network of Evidence in the Wojciechowski Study

AVA = avatrombopag; ELT = eltrombopag; FOS = fostamatinib; PLC = placebo; ROM = romiplostim.

Source: Wojciechowski et al.⁷ This work is licensed under the Creative Commons Attribution 4.0 International (CC BY-NC 4.0) Licence. Full text available at https://creativecommons.org/licenses/by-nc/4.0/.

In this ITC, the number of patients (12 to 135 in active treatment arms), length of follow-up (24 to 36 weeks), and median duration of disease (1.6 to 10.8 years) varied across the trials. The median age and median platelet count at baseline ranged from 41 to 57 years and 14 × 10⁹/L to 24 × 10⁹/L, respectively. The number of previous treatments and concomitant ITP therapy also varied across trials (Table 18).

Data were available allowing a comparison of the following outcomes for all treatments: durable platelet response, need for rescue treatment, and WHO grade 2 to 4 bleeding events. Data for reduction in the use of concomitant ITP therapies were not available for fostamatinib. The definitions of durable response were relatively similar for all the treatments to allow for comparisons, although the definition was different in the FIT1 and FIT2 trials (4 of 6 visits over weeks 14 to 24) compared with the other trials (at least 6 of the last 8 weeks of treatment). To compare bleeding events among all treatments, it was assumed that WHO grade 2 to 4 bleeds were equivalent to grade 2 to 5 bleeds reported in NCT00102336 and moderate to severe bleeds reported in the FIT1 and FIT2 studies.

Results

Sponsor-Submitted Indirect Treatment Comparison

Overall platelet response: The definition of overall platelet response varied across studies included in the ITC. A series of analyses was performed using different definitions of platelet response. Results of analyses 1, 2, and 3 are presented in Table 19. The difference between fostamatinib and rituximab was statistically significant in analysis 1 only.

Wojciechowski Study

Durable platelet response: Based on evidence from 6 studies, there was no difference between fostamatinib and various TPO-RAs in durable platelet response.

Reduction in use of concomitant ITP medication: Data for reduction in the use of concomitant ITP therapies were not available for fostamatinib.

Need for rescue therapy: Based on 6 studies, there was no difference between fostamatinib and TPO-RAs in the need for rescue therapy.

Incidence of any bleeding events: Based on results from 7 studies, there was no difference between fostamatinib and TOP-RAs in the incidence of any bleeding events.

Incidence of WHO grade 2 to 4 bleeding events: Based on 6 studies, there was no difference between fostamatinib and TPO-RAs in the incidence of WHO grade 2 to 4 bleeding events.

Any adverse event: Based on the results from 5 studies, there was no difference between fostamatinib and TPO-RAs in the incidence of any AEs.

Details of the results are presented in Table 20.

Critical Appraisal of the Sponsor-Submitted Indirect Treatment Comparison and Wojciechowski Study

In both ITCs, the analysis of efficacy and safety data presented was limited by the size of the evidence base.

Sponsor-Submitted Indirect Treatment Comparison

In this ITC, studies were identified and selected using a systematic review approach. For example, multiple databases were searched, and 2 independent reviewers selected the studies based on specific inclusion and exclusion criteria and performed data extraction. Quality assessment of the included studies was not performed.

One outcome (overall platelet response) was evaluated in this study. Because there is no clear evidence of a direct correlation between the degree of thrombocytopenia and bleeding symptoms, it is unclear whether treatment with fostamatinib would be useful in improving clinical outcomes such as reduction in subsequent bleeding events and the need for rescue therapy, and improvement in patient HRQoL.

Trial characteristics and patient baseline characteristics of the studies included in the systematic review and ITC were reported. Based on the data presented, potential sources of heterogeneity with respect to the baseline characteristics were identified, such as disease duration (ranging from 24 to 78 weeks) and definition of overall platelet response. During the feasibility analysis period, several potential treatment-effect modifiers were identified by the sponsor, such as baseline demographic characteristics, medical history (e.g., time since ITP diagnosis, prior treatment for ITP, and concomitant medications). However, the clinical experts consulted by the sponsor indicated that none of these patient characteristics could be considered treatment-effect modifiers in the study population (patients with chronic and persistent ITP who can receive treatment of fostamatinib or rituximab) due to a lack of evidence. According to the clinical experts consulted by CADTH, these are all important treatment-effect modifiers, as are other patient characteristics in the study population, such as cycles and doses of prior corticosteroid therapy, previous lines of therapy, and severity of previous bleeding events. However, such data were not provided in the ITC, and heterogeneity across the included trials needs to be further assessed and adjusted.

Because the definition of platelet response varied across included trials, sensitivity analyses based on different definitions of this outcome were performed. Results of these analyses were consistent and favoured fostamatinib, showing that treatment with fostamatinib was associated with a higher probability of achieving an overall platelet response compared with rituximab, although the difference was statistically significant only when the definition of overall platelet response from each study was used.

A narrower patient group was the interest of this ITC. The authors assumed that patients who were successfully treated with a splenectomy or TPO-RAs were unlikely to require additional treatment with a new intervention, and they stated that the target population of this study was patients who did not have access to long-term effective therapy options, such as those who received short courses of rituximab and those on a watch-and-rescue regimen. Rituximab was therefore the only comparator in this ITC, and its study results can be generalized only to patients with persistent or chronic ITP who did not receive prior TPO-RA therapy or had not undergone a splenectomy. The absence of other comparators is a key limitation of this ITC.

Wojciechowski Study

In this ITC, studies were identified and selected according to pre-specified inclusion and exclusion criteria. Multiple databases were searched. The authors stated that the study was conducted in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, which suggested that the standard practice of minimizing the bias of selecting studies for inclusion had been used. Quality assessment of the included studies was performed using a validated tool. However, there was no discussion of how any potential biases in the trials could affect data analyses in the ITC and possible solutions (e.g., whether sensitivity analyses were conducted to assess the impact of studies with poor quality). Multiple clinical outcomes, including the incidence of AEs, were evaluated in this study, which allows for a comprehensive evaluation of the clinical benefits and risks of the study drugs. Definitions of these outcomes were similar across the trials.

Trial characteristics and patient’s baseline characteristics of the studies included in the systematic review and ITC were reported. Potential sources of heterogeneity with respect to the baseline characteristics were identified based on these data, such as disease duration (ranging from 1.6 to 10.8 years), number of previous treatments, and concomitant ITP medication. This difference in patient’s baseline characteristics may vary the response between groups and may not allow groups to be comparable.

The analysis of efficacy and safety data presented was limited by the size of the evidence base. Due to the small evidence base and low incidence of SAEs across all studies, the results of this analysis are largely noninformative. This is particularly true when looking at comparative safety issues, especially for inclusion in economic models and comparative efficacy studies, resulting in imprecision due to small effect sizes and large CrIs.

Summary

Based on the results of the sponsor-submitted ITC, fostamatinib was favoured over rituximab in achieving overall platelet response in patients with persistent or chronic ITP. However, this study has a number of limitations that affect internal and external validity, such the inability to comprehensively assess clinical heterogeneities across the included studies and their impact on the study results, the inability to explore the relative treatment effect of fostamatinib in various subgroups, and the lack of other important efficacy and safety data for the ITP treatments of interest. In addition, the comparative efficacy and safety of fostamatinib to other active ITP therapies, such as TPO-RAs, was unknown for the study population due to a lack of evidence.

Based on the findings from the Wojciechowski ITC, which evaluated the relative efficacy and safety of currently available active treatments for patients with chronic ITP who did not have adequate response to corticosteroids, it remains uncertain if there was a significant difference between fostamatinib and TPO-RAs in achieving durable platelet response, reducing the need for rescue therapy, and decreasing bleeding and other AEs. The applicability of this ITC is affected by the limited size of the evidence base and heterogeneity in patient populations across trials. Overall, the results of this analysis must be interpreted with caution.

Other Relevant Evidence

This section includes an open-label extension study and 2 additional relevant studies included in the sponsor’s submission to CADTH that were considered to address important gaps in the evidence included in the systematic review.

Long-Term Extension Studies

One open-label extension study (FIT3) has been summarized to provide additional evidence regarding the long-term safety and efficacy of fostamatinib 150 mg twice daily if the platelet count was less than 50,000/µL and the study drug was well tolerated or reduced to a dose as low as fostamatinib 100 mg once daily if dose-limiting AEs (as defined by the protocol for patients with persistent or chronic ITP) were observed. Data for this summary were presented in the Clinical Study Report dated October 15, 2014, with a data cut-off for the final report of June 01, 2020.⁹

Methods

The FIT3 (C788 to 049) trial was a multi-centre, phase III, open-label extension study conducted at 54 sites in 16 countries (Canada, the US, Australia, the European Union, and the UK). The primary objective was the assessment of the long-term safety of fostamatinib among patients with persistent or chronic ITP. The secondary objectives were the establishment of the long-term efficacy of the drug as well as the assessment of the pharmacokinetics profile.

The trial consisted of 18 monthly visits followed by every-other-month visits for a maximum of 5 years of treatment or until the drug was commercially available for all patients, whichever occurred first. A total of 123 patients from the FIT1 (C788 to 047) and FIT2 (C788 to 048) trials who completed the week 24 evaluation or withdrew early (starting at week 12) due to a lack of response were included in this trial. Patients were blinded to their respective treatment assignment, either active or placebo, in the FIT1 or FIT2 studies.

During the FIT3 trial, all patients received open-label fostamatinib. Moreover, patients could continue to receive concomitant ITP medications that were allowed in the FIT1 and FIT2 studies or could receive a reduced dose of the concomitant medications in case their platelet count was stable at 50,000/µL or greater. Patients were allocated into 2 treatment groups, responders and nonresponders, depending on their response in the previous FIT1 or FIT2 studies. The responders group (last platelet count ≥ 50,000/µL) initiated open-label fostamatinib treatment using the same dosage and regimen (150 mg twice a day or 100 mg twice a day) that achieved a stable platelet count in the previous study, whereas the nonresponders group (last platelet count < 50,000/µL) initiated their treatment with 100 mg twice a day during the trial (Figure 5). At month 1, the dosage for patients showcasing a platelet count of less than 50,000/µL and tolerating the study drug well was increased to 150 mg twice a day. The dosage of fostamatinib was reduced to as low as 100 mg once daily if any dose-limiting AEs were observed among patients.

Figure 5: Initial Treatment Allocation

Note: Doses were administered twice a day unless the dose was reduced to once a day.

* Treatment assignment and dose from the previous study.

** At month 1, patients receiving fostamatinib 100 mg twice a day had the dose escalated to fostamatinib 150 mg twice a day if the platelet count was less than 50,000/µL and fostamatinib was well tolerated.

Source: Clinical Study Report for FIT3.⁹

After 12 weeks (3 months) of treatment, as well as 4 or more weeks at a dosage of 150 mg twice a day, patients whose platelet count was consistently less than 50,000/µL were required to withdraw from the trial.

Populations

Patients were eligible to participate in the FIT3 study if they met the following inclusion criteria:

• They were able to give written informed consent.

• They completed the week 24 evaluation of FIT1 or FIT2 studies or withdrew early (starting at week 12) due to a lack of response, with up to 7 days between the last day of treatment in previous study and initial dosing in FIT3.

• They were male or female, aged at least 18 years.

• Female patients were either postmenopausal for at least 1 year or surgically sterile; in the case of childbearing potential, they were not pregnant or lactating and using birth control method during the whole study period and for 30 days after the last dose.

• They had the ability to understand the nature of the study and associated hazards, as well as to communicate effectively with the investigator regarding these issues.

Patients were not eligible to participate in the FIT3 study if they met any of the following exclusion criteria:

• They had withdrawn from the FIT1 or FIT2 study before week 12 or for any reason except lack of response.

• They had poorly controlled hypertension (persistent or repeated systolic ≥ 140 mm Hg or diastolic ≥ 90 mm Hg) during FIT1 or FIT2, irrespective of the receival of anti-hypertensive treatment.

• They had laboratory abnormalities during enrolment, such as a neutrophil count of less than 1000/µL, a leukocyte count of less than 2000/µL, a lymphocyte count below 750/µL, hemoglobin below 10 g/dL, or transaminase levels (ALT or AST) above 1.5 × the upper limit of normal, bilirubin > 1.5 × the upper limit of normal, or an estimated glomerular filtration rate below 30 mL/minute.

• They had significant infection, acute infection such as influenza, or some known inflammatory process.

• They had received any blood or blood products within 2 weeks before their enrolment. (Exceptions could be IVIG or IV anti-D immunoglobulin, if used as rescue therapy.)

Patient baseline characteristics are summarized in Table 21. Among patients enrolled in the FIT3 trial, the mean age was 51.8 years (SD = 15.9), more than half of the patients were female (60.2%), and the mean body mass index was 28.37 kg/m² (SD = 7.66).

Interventions

During the trial, all patients received open-label fostamatinib, which was self-administered twice a day by mouth, once in the morning and once in the evening, for up to 5 years or until the drug became commercially available for all patients, whichever happened first. In cases of reduced doses due to AEs, the drug had to be taken in the morning. In the event of a missing dose, patients were advised to take the next dose and not to take 2 doses simultaneously.

The initial treatment allocation is shown in Figure 5 . Patients were assigned to 1 of 2 groups of treatments, responders and nonresponders, depending on their response in the previous FIT1 or FIT2 studies. The responders group (last platelet count ≥ 50,000/µL) initiated open-label fostamatinib treatment with the same dosage and regimen (150 mg twice a day or 100 mg twice a day) that achieved a stable platelet count in the previous study, whereas the nonresponders (last platelet count < 50,000/µL) initiated their treatment with 100 mg twice a day during the trial. At month 1, the dosage for patients showing a platelet count of less than 50,000/µL and tolerating the study drug well was increased to 150 mg twice a day. The dosage of fostamatinib was reduced to as low as 100 mg once daily if any dose-limiting AEs were observed among patients.

During the trial patients were allowed to continue any concomitant medications and/or treatments for ITP therapies that were allowed in the FIT1 and FIT2 studies. Allowed concomitant medications and/or treatments for ITP therapies were glucocorticoids at a dosage of up to 20 mg of prednisone, danazol, or azathioprine daily. For patients who demonstrated a stable platelet count of 50,000/µL or greater, tapering of the dose of their concomitant medication was considered. No new treatments for ITP were allowed except as rescue medication. Because it was an open-label extension, all patients, investigators, and staff members were aware of the treatment received, which was fostamatinib. However, each patient remained blinded to their assignment group from the previous FIT1 and FIT2 studies during trial.

Outcomes

The primary objective of the FIT3 trial was the assessment of the long-term safety of fostamatinib among patients with persistent or chronic ITP. The secondary objectives were the establishment of the long-term efficacy of the drug as well as the assessment of the pharmacokinetic profile. The primary efficacy outcome of fostamatinib were summarized in this report by the platelet counts and the requirement of dose adjustment. Stable platelet response was a platelet count of 50,000/μL or greater at 4 or more of 6 biweekly visits during weeks 14 to 24 or, for patients initiating fostamatinib in the extension phase, at least 1 platelet count of at least 50,000/μL in the first 3 months of fostamatinib treatment followed by platelet counts of at least 50,000/μL at the subsequent 2 of 3 monthly visits without use of rescue medication. The primary efficacy outcome had 2 versions. For version 1, efficacy was assessed among patients who were on active treatment in either the FIT1 or FIT2 studies, in the current extension study, or in both. For version 2, efficacy was assessed among patients assigned to placebo in either of the prior FIT1 or FIT2 studies. The secondary efficacy outcomes were reported as the duration of platelet response among patients and the response (yes or no) among patients with a reduction in the dose of concomitant ITP medication while maintaining an adequate platelet count. For the safety measurement, the outcomes assessed and summarized in the report were the frequency and severity of bleeding according to the IBLS and WHO bleeding scales; change from baseline in liver function, blood pressure, and neutrophil count; the incidence and severity of gastrointestinal effects (diarrhea, nausea, vomiting, and abdominal pain), infection, and overall AEs. For the pharmacokinetic end point summary, the profile of R406 (the active metabolite of fostamatinib), was determined, including maximum plasma concentration, area under the curve, and time to maximum plasma concentration in a subset of patients.

Statistical Analysis

Sample size was calculated using the patients’ numbers in the prior FIT1 and FIT2 studies and their eligibility and willingness to participate in the FIT3 trial. The efficacy and safety analyses were performed among the treated population, including all enrolled and treated patients in this trial. The primary efficacy measurement was calculation of a platelet count at every study visit. The primary efficacy end point was analyzed in 2 versions.

Version 1 included the assessment of efficacy among all patients while they were on active treatment in 1 of the prior FIT1 or FIT2 studies, in the present FIT3 extension study, or in both. For this version, the achievement and maintenance of a stable platelet count was defined as achievement of a platelet count of 50,000/μL or greater within 12 weeks of beginning active treatment; and achievement of a sustained stable platelet response, defined as no 2 visits, at least 4 weeks apart, with a platelet count below 50,000/μL and no intervening visit with a platelet count of 50,000/μL or greater unrelated to rescue therapy, within a period of 12 months following initial achievement of the target platelet count.

Version 2 was the assessment of a within-patient, between-study comparison of fostamatinib and placebo among patients assigned to placebo in either of the prior FIT1 or FIT2 studies. For this version the achievement and maintenance of a stable platelet count was defined as achievement of a platelet count of 50,000/μL or greater within 12 weeks of beginning treatment (placebo treatment in the prior FIT1 or FIT2 study and fostamatinib treatment in the FIT3 extension study); and achievement of a sustained stable platelet response, defined as no 2 visits, at least 4 weeks apart, with a platelet count below 50,000/μL and no intervening visit with a platelet count of 50,000/μL or greater, unrelated to rescue therapy, within a period of 12 weeks following initial achievement of the target platelet count.

For the assessment of secondary efficacy outcome, the Kaplan–Meier method was used to analyze the duration of platelet response. Descriptive statistics were used to perform and summarize additional platelet counts analyses.

For the safety analysis, coding of AEs according to the Medical Dictionary for Regulatory Activities version 18.1 was used. Descriptive statistics were presented by visit for the actual values and the changes from baseline for vital signs and quantitative laboratory tests. The mean of the IBLS scores across visits during the treatment period was also summarized using descriptive statistics.

Patient Disposition

Patient disposition is summarized in Table 22. A total of 124 patients out of 150 from the prior FIT1 and FIT2 studies were screened for the FIT3 trial. Of those 124 individuals, 1 failed screening and 123 patients were enrolled. Among these 123 patients in the treated population, 44 were randomized to placebo and 79 to fostamatinib in their previous study (FIT1 or FIT2). Among the 123 patients, 94 (76.4%) withdrew from the study prematurely. The most common reason for patient withdrawal (accounting for 35.9% of all patients) during this trial was a lack of platelet response after week 12 or later. Other reasons for withdrawal included patient’s own decision for 11 patients (8.9%), development of AEs related to the trial for 11 (8.9%), other AEs for 10 (8.1%), lack of a platelet response before week 12 for 9 (7.3%), sponsor’s decision for 6 (4.9%), noncompliance for 2 (1.6%), and investigator’s discretion for 1 (0.8%) patient for being nonresponsive to fostamatinib.

Exposure to Study Treatments

The mean duration of exposure to fostamatinib was 598.7 days (SD = 607.2) and the median duration of exposure to fostamatinib was 180.0 (range = 28 to 1,773) days.

Some therapeutic treatments were allowed as rescue therapy for patients experiencing platelet counts of less than 50,000/µL. The rescue therapy was administered to the patients in the following situations:

• a platelet count of less than 50,000/µL and at immediate risk of bleeding or with clinically significant bleeding or wet purpura

• a platelet count of less than 50,000/µL and requiring urgent or emergent surgery.

For the FIT3 trial, the allowed therapeutic regimens included:

• IVIG: up to 1 g/kg × 1 to 3 days, or

• IV anti-D immunoglobin G: up to 50 to 75 mcg/kg × 1 to 2 days, or

• IV methylprednisolone up to 1 g/day for 1 to 3 days or oral dexamethasone up to 40 mg/day for 1 to 2 days or oral prednisone up to 1 mg/kg/day for 1 to 3 days.

Treatment compliance was calculated as the ratio of the total number of actual doses divided by the total number of expected doses during the treatment period in the FIT3 trial. The mean overall compliance was 104.79% (SD = 29.52) for fostamatinib, the mean overall compliance was 99.5%, and the median (number of missed doses) was 1.0 (range = 0 to 803).

Efficacy

Efficacy results are summarized in Table 23. For the primary efficacy outcome (version 1), 19 patients (15.4%) had a platelet response within 12 weeks of taking fostamatinib and maintained a stable platelet response for at least 12 months after achieving the initial response (95% CI, 9.6% to 23.1%). For the primary efficacy outcome (version 2), among 44 patients who were treated with placebo in the FIT1 or FIT2 trials and fostamatinib in the FIT3 trial, 10 (22.7%) were responders, while 34 (77.3%) remained nonresponders in both the prior trials and the FIT3 trial.

For the secondary efficacy outcome, 13 of the 19 primary efficacy responders (version 1) who were treated with fostamatinib in 1 of the prior studies, the minimum and maximum duration of response was 427 days and 1,661 days, respectively. In 6 of the 19 responders who received placebo in 1 of the prior studies (version 1), the minimum and maximum duration of response was 1,340 days and 1,743 days, respectively. In 10 of the 19 responders who got placebo in 1 of the prior studies (version 2), the minimum and maximum duration of response was 194 days and 1,743 days, respectively.

Harms

Safety results are summarized in Table 24. Most patients (79.7%) experienced at least 1 AE during the treatment phase of the FIT3 trial. The most frequently reported AEs were diarrhea (29.3%), hypertension (17.9%), petechiae (15.4%), epistaxis (15.4%), headache (12.2%), upper respiratory tract infection (11.4%), dizziness (10.6%), contusion (9.8%), nausea (8.9%), vomiting (8.9%), fatigue (8.1%), cough (8.1%), and thrombocytopenia (8.1%). Serious AEs with were reported for 27.6% of patients, with thrombocytopenia being the most frequently reported among 6.5% patients.

During the FIT3 trial, a total of 18 patients (14.6%) withdrew from the trial due to an AE, with diarrhea being the most commonly reported AE among 4.1% patients, followed by neutropenia and increased hepatic enzyme among 1.6% patients. Four people (3.3%) died, 1 each due to the following AEs: endocarditis bacterial, pneumonia, sepsis, and cardio-respiratory arrest.

Critical Appraisal

Internal Validity

The open-label extension study had several limitations imposed by the overall design. First, the lack of a randomized comparison group to provide context and control for potential confounders and the open-label design may influence the perception of improvement by patients and clinicians, which could affect the reporting of harms. As part of the eligibility criteria for the open-label extension, patients had to complete 1 of the prior studies, potentially allowing for selection bias. From the total population that completed the FIT1 and FIT2 studies (n = 150), a total of 124 (82.7%) enrolled in the open-label extension. There was no clear explanation for half on those 26 patients (17.3%) not enrolling in the FIT3 trial. Additionally, there was a potential for survival bias as the other 13 patients who discontinued the prior studies due to AEs were excluded (7 in the fostamatinib group and 2 in the placebo group in the FIT1 trial, and 2 in the fostamatinib group and 2 in the placebo group in the FIT2 trial). This could result in a greater enrolment of patients who were better able to tolerate fostamatinib and possibly fewer AEs being reported. Any lack of follow-up after discontinuing FIT3 could mean that important long-term safety data are also missing. The high rate of discontinuations (76.4%) during the open-label phase was mostly due to a lack of response, with the most common reasons being a lack of platelet response at week 12 or later (35.8%), patient decision (8.9%), development of a study-specific AE (8.9%), and other AEs (8.1%). The limitations with the study design make it challenging to interpret the results and form conclusions with certainty.

External Validity

Because the patients who took part in the FIT3 were originally from the parent studies and the eligibility criteria remained the same, it is reasonable to expect that the same limitations to generalizability are relevant to the open-label extension. For example, because the participants were predominantly White (91.9%), the results from these trials may not be generalizable to other racial groups commonly seen at some centres in Canada (although the experts noted that the role of race and/or ethnicity in treatment response was not clear). The experts also noted that patients with secondary ITP were excluded from the FIT3 trial, and trial findings may therefore not be generalizable to those with secondary ITP. Moreover, the experts found the definition for response to therapy in the FIT3 trials, as in the FIT1 and FIT2 trials, to be overly strict, and they stated that achieving a platelet count of at least 30,000/µL would be considered a response in practice, particularly if a patient is asymptomatic. The primary outcome definition used in the FIT3 trial therefore may not be reflective of how response to therapy is assessed in routine practice. The clinical experts also noted that the co-interventions (i.e., concomitant ITP medication) used in the FIT3 trial, as in the FIT1 and FIT2 trials, are also reflective of real-world practice in Canada, as steroids would commonly be continued while a patient is initiated on a new long-term treatment strategy such as fostamatinib.

As with the FIT1 and FIT2 trials, the FIT3 trial provided limited data on clinically important outcomes such as quality of life, rescue therapy, and bleeding events. Because the clinical experts do not use the IBLS and WHO bleeding scales in practice, the relevance of the bleeding outcome scales used in the FIT3 trial is unclear. Further, the event rates for the post hoc bleeding-related SAE outcome made it challenging for the clinical experts to comment on the relevance or meaningfulness of these findings.

Post Hoc Analyses of the FIT1, FIT2, and FIT3 Trials

In a post hoc analysis conducted by Boccia et al. (2020),²⁶ the efficacy of fostamatinib was assessed earlier in the ITP disease course by including patients from the FIT1, FIT2, and FIT3 studies. The data cut-off date for this post hoc analysis was December 2019. For the evaluation process, patient subgroups were compared by line of therapy (second-line versus third- or later-line) and the chronic ITP progression stage (persistent versus early or late stage of disease). The proportions of patients achieving a platelet response of 50,000/µL or greater and 30,000/µL or greater, without any rescue therapy within 4 weeks, were assessed. A total of 145 patients were assessed in this study, 32 receiving fostamatinib as second-line treatment (median age of 50; 59% female), and 113 as other-line treatment (median age of 54; 60% female). The response rates were observed to be higher in the second-line therapy group (78% using fostamatinib) compared with the third or later lines of therapy (48%), presenting fostamatinib as more effective as second-line therapy for chronic ITP patients. The safety results were similar for patients in both therapy lines.²⁶ The authors reported that AE rates were 72% for second-line therapy and 94% for later-line therapy. The most common AEs were hypertension (31% in second-line versus 19% in later-line therapy), diarrhea (25% versus 39%, respectively), upper respiratory tract infections (16% versus 11%, respectively), and elevated liver transaminase (26% versus 16%, respectively). Considering the limitations of the post hoc analysis, the results from this study should be interpreted with caution. As this study included patients from the FIT1, FIT2 and FIT3 studies, the corresponding internal and external validity considerations would be applicable to this post hoc analysis as well. In addition, lack of randomization or stratification for lines of therapies before fostamatinib treatment and the small number of patients in the second-line therapy group compared to the third- and other-line of therapy may potentially lead to bias. The differences in the 2 groups in terms of time since ITP diagnosis, stages of ITP progression, baseline platelet counts, and prior ITP medications used may have affected the efficacy of fostamatinib in those groups differently.

In another post hoc analysis conducted by Cooper et al. (2021),²⁷ thrombotic risk was assessed while patients were on long-term treatment with fostamatinib. A total of 146 patients from the FIT1, FIT2, and FIT3 trials were included in this study, with up to 5 years of treatment and a data cut-off date of December 2019.^23,28 Although the investigators claimed an effective and long-term treatment for ITP with a low incidence of thromboembolic events among patients, the results should be interpreted with caution considering the post hoc nature of the trial and the prospect of associated bias.

Discussion

Summary of Available Evidence

The FIT1 (N = 76) and FIT2 (N = 74) trials were identically designed 24-week double-blind RCTs that evaluated the efficacy and safety of fostamatinib versus placebo in patients with primary ITP for longer than 3 months who had received at least 1 previous ITP treatment and had a baseline platelet count below 30,000/µL. In the FIT1 trial, 51 patients were randomized to fostamatinib and 25 to placebo, while in the FIT2 trial, 50 patients were randomized to fostamatinib and 24 to placebo. The primary efficacy end point in both trials was achievement of a stable platelet response, defined as a platelet count of 50,000/µL or greater at 4 of the last 6 study visits between weeks 14 and 24.

Two ITCs were reviewed. The Wojciechowski study was a systematic review and ITC comparing fostamatinib to 3 TPO-RAs (avatrombopag, eltrombopag, and romiplostim) among patients with chronic ITP who had inadequate response to previous therapy. Seven phase III, double-blind RCTs were included and contributed data on various efficacy and harms outcomes to the ITC. The authors assessed the following outcomes: durable platelet response, need for rescue therapy, and WHO bleeding events, all up to 24 weeks. The sponsor also submitted a systematic review and ITC in which fostamatinib was compared to rituximab among patients with chronic or persistent ITP. Six RCTs were included and contributed evidence. A single outcome, overall platelet response, was assessed in this ITC.

One additional study, FIT3, was considered as other relevant evidence. This was an open-label extension study of FIT1 and FIT2 to examine the efficacy and safety of long-term fostamatinib among people with chronic or persistent ITP. A total of 59 patients from FIT1 and 64 patients from FIT2 who completed the week 24 evaluation or withdrew early (starting at week 12) due to a lack of response were eligible for this trial. All patients received open-label fostamatinib. The primary efficacy outcome was achievement of a platelet response by 12 weeks and maintenance for 12 months. A stable platelet response was a platelet count of 50,000/μL or greater at 4 or more of 6 biweekly visits during weeks 14 to 24 or, for patients initiating fostamatinib in the extension phase, a platelet count of at least 50,000/μL in the first 3 months followed by platelet counts of at least 50,000/μL at the subsequent 2 of 3 monthly visits without use of rescue medication. In the FIT3 trial, 60% of patients were female, the mean age was 52 years (SD = 16), and patients were predominantly White (92%).

A post hoc analysis of the FIT1, FIT2, and FIT3 trials was conducted by Boccia et al.²⁶ and was also considered relevant. Patient subgroups were compared by line of therapy for fostamatinib (second-line versus third- or later-line) and the chronic ITP progression stage (persistent versus early or late stage of disease). A total of 145 patients were assessed in this study, 32 receiving fostamatinib as second-line treatment and 113 as other-line treatment.

Interpretation of Results

Efficacy

The results of the FIT1 and FIT2 trials suggest that fostamatinib leads to a modestly higher rate of stable treatment response (approximately 18%) compared to placebo (0 to 4%) in a group of heavily pre-treated patients with primary, chronic ITP. While the use of rescue therapy and rate of bleeding-related SAEs were lower in the fostamatinib groups compared to placebo, the event rates for these outcomes were low and the trials were not powered to detect differences in these outcomes. Further, due to the high discontinuation rate in the FIT1 and FIT2 trials, limited data were available on quality of life (measured by the SF-36), and it was impossible to determine whether there were any differences between fostamatinib and placebo. Subgroup analyses suggested that there were no differences in platelet response for people treated with a prior splenectomy or TPO-RAs versus those who were not so treated, but these analyses were not pre-specified and were likely underpowered (with wide CIs and low event rates). The clinical experts consulted by CADTH noted that it would be challenging to base treatment decisions on, or draw meaningful conclusions from, the subgroup analyses. Overall, the clinical experts emphasized that another treatment option that produces a modest improvement in platelet response would be helpful in clinical practice, given that patients are often refractory to multiple treatments.

In the FIT3 trial, long-term platelet response (≥ 12 months) was observed in 15% of patients who responded in the FIT1 and FIT2 trials, and 23% of patients of patients receiving placebo in the FIT1 and FIT2 trials but starting fostamatinib in FIT3. These results suggest that a portion of patients will achieve long-term platelet responses; however, there was no comparator group in this extension study, it was open-label, and there was a potential for selection bias. The FIT1 and FIT2 trials were conducted in a primarily White population, and patients with secondary ITP were excluded. In addition, the rates of prior treatments (e.g., rituximab and a splenectomy) may not reflect Canadian practice, creating generalizability concerns in the Canadian context.

The clinical experts noted a lack of comparative efficacy data for second- and third-line ITP treatments, including with fostamatinib. Two ITCs provided evidence on comparative efficacy. Wojciechowski et al. reported that no treatment was favoured when fostamatinib was compared to TPO-RAs among patients who did not achieve an adequate response to corticosteroids in terms of platelet response rates and reduction in bleeding event outcomes, and it remains uncertain if there was a significant difference for fostamatinib compared to TPO-RAs in achieving a durable platelet response. There were important limitations in this ITC due to the small size of the evidence base and heterogeneity in patient populations across included trials. The sponsor also submitted an ITC that suggested that fostamatinib was superior to rituximab in achieving an overall platelet response among patients with persistent or chronic ITP. This ITC also was not able assess clinical heterogeneities across eligible studies or their impact on study results. Moreover, the study only assessed platelet response and had a potential for selective reporting. Due to limited comparators and important methodological limitations, the ITCs provide limited additional insight into comparative efficacy of fostamatinib compared to other second- or subsequent-line ITP therapies.

The lack of comparative efficacy and safety data, as well as uncertainty around the optimal treatment pathway in second- and subsequent-line treatment of ITP, is reflected in guidelines for ITP.^2,3 These guidelines highlight the low certainty of evidence regarding ITP treatment options, making it difficult to weigh options against each another. Guidelines therefore acknowledge that individualization of therapy and shared decision-making are important in the treatment of ITP and should incorporate the duration of ITP, comorbidities, age of the patient, access to medications (cost and availability), and patient preferences.³ This consideration was echoed by clinical experts, who noted that it was challenging to compare fostamatinib to other second- or subsequent-line ITP treatment options but that the modest efficacy with respect to platelet count response meant it represented another treatment option among heavily pre-treated patients who are in need of options for managing ITP.

Clinical experts and patients highlighted how reducing bleeding risk and improving symptoms and quality of life are particularly important in chronic ITP, considering how these factors negatively affect patients. The patient group highlighted how symptom and quality-of-life improvement is likely more important to patients compared with platelet counts. Clinicians and patients also reported that an ideal ITP treatment should be convenient and easy to administer for a patient. Unfortunately, the available evidence provided limited insight on outcomes important to patients and clinicians as the eligible and relevant trials (and ITCs) focused primarily on platelet counts. Some outcomes important to patients, such as bleeding and quality of life, were secondary or post hoc outcomes with low event rates and/or limited outcome data, and it was not possible to draw conclusions about fostamatinib’s effect on these outcomes. While fostamatinib is an oral medication, convenience and adherence were not compared in any of the relevant evidence and as such the extent to which fostamatinib leads to improvements in these measures compared to existing treatments is unclear. Platelet response is the main way that clinicians assess treatment response in clinical practice. The clinical experts noted that platelet response is expected to correlate with reduced bleeding risk; however, they acknowledged that available evidence with fostamatinib provides limited insight regarding its effect on quality of life, symptoms, and bleeding outcomes (bearing in mind that there are limited data on such outcomes for most ITP treatments).

Harms

In the FIT1, FIT2, and FIT3 trials, fostamatinib was generally well tolerated. The most common AEs were diarrhea, nausea, elevated liver transaminase levels, and hypertension. The rate of infections in the FIT1 trial was higher in the fostamatinib group (|||) compared to placebo (|||), which was mainly attributed to upper respiratory tract infections and urinary tract infections. The rate of SAEs was similar between groups in the FIT1 trial, while in the FIT2 trial the placebo group had a higher rate of SAEs (26% versus 10% in fostamatinib group). The rate of withdrawals due to AEs was higher in the fostamatinib group compared to placebo in the FIT1 trial (16% versus 8%, respectively), and the rate of withdrawals due to AEs in both groups was similar in the FIT2 trial. The FIT3 trial did not identify any long-term safety concerns with fostamatinib, although there was no comparator group in this study. The clinical experts did not highlight any major concerns regarding the safety of fostamatinib, noting that gastrointestinal adverse effects, hypertension, and elevated transaminase levels were consistent with what is known about fostamatinib. They noted that hypertension would be important to monitor in patients treated with fostamatinib. The clinician group input stated that fostamatinib is considered to have a favourable side-effect profile in the context of ITP treatment. The FIT1, FIT2, and FIT3 trials did not provide insight on the comparative safety of fostamatinib versus other second- or subsequent-line therapies for ITP. The Wojciechowski ITC found no statistically significant differences between fostamatinib and TPO-RAs in the incidence of any AEs; however, this ITC had important limitations. The data on the safety and tolerability of fostamatinib compared to other ITP treatments are therefore limited.

Conclusions

Management of chronic ITP is challenging as patients frequently relapse or are refractory to treatments, and patients often cycle through multiple ITP treatments. Treatment is complicated by a lack of evidence on comparative efficacy and the safety of second- and subsequent-line treatment options, access issues, and the safety and tolerability of available options. In 2 double-blind RCTs, fostamatinib, which is an ITP treatment with a novel mechanism of action, led to a modest improvement in platelet count responses compared to placebo among patients with chronic, heavily pre-treated primary ITP. There were little or no data on outcomes important to patients, such as bleeding rates, symptoms, and quality of life, and the impact of fostamatinib on these outcomes remains unclear. Subgroup analyses (e.g., based on previous lines of therapy) were not able to provide insight into which patient groups are the most likely to respond to treatment. It is also difficult to draw conclusions about the comparative efficacy of fostamatinib versus other ITP treatments. Both ITC studies included in this review suggest that fostamatinib may be comparable to TPO-RAs and have a favourable efficacy comparable to that of rituximab in terms of platelet count response. However, there were important limitations in these studies, and it is challenging to draw firm conclusions about comparative efficacy based on these studies. In the FIT1 and FIT2 trials, fostamatinib appeared to lead to a higher rate of adverse effects such as diarrhea, nausea, hypertension, and elevated liver transaminase levels, compared to placebo, while the FIT3 trial did not identify any long-term safety concerns beyond these adverse effects. Overall, this review suggests that fostamatinib is another potential treatment option for patients with chronic, heavily pre-treated primary ITP. The drug leads to a platelet count response in a modest proportion of patients and is generally well tolerated compared to placebo, although its comparative efficacy and safety versus other ITP treatments, and its effect on outcomes important to patients, remain unclear.

References

1.Arnold DM, Leung LKL, Tirnauer JS. Immune thrombocytopenia (ITP) in adults: clinical manifestations and diagnosis. In: Post TW, ed. UpToDate. Waltham (MA): UpToDate; 2021: www.uptodate.com. Accessed 2021 Oct 20.

2.Provan D, Arnold D, Bussel J, et al. Updated international consensus report on the investigation and management of primary immune thrombocytopenia. Blood Adv. 2019;3(22). PubMed

3.Neunert C, Terrell D, Arnold D, et al. American Society of Hematology 2019 guidelines for immune thrombocytopenia. Blood Adv. 2019;3(23). PubMed

4.Tavalisse (fostamatinib): 100 mg and 150 mg tablet (as as fostamatinib disodium hexahydrate), oral [product monograph]. Toronto (ON): Medison Pharma Canada Inc.; 2020 Nov 17.

5.Clinical Study Report: C-935788-047. A phase 3, multicenter, randomized, double-blind, placebo-controlled study of fostamatinib disodium in the treatment of persistent/chronic immune thrombocytopenic purpura [internal sponsor's report]. San Francisco (CA): Rigel Pharmaceuticals, Inc.; 2017 Mar 3.

6.Clinical Study Report: C-935788-048. A phase 3, multicenter, randomized, double-blind, placebo-controlled study of fostamatinib disodium in the treatment of persistent/chronic immune thrombocytopenic purpura [internal sponsor's report]. San Francisco (CA): Rigel Pharmaceuticals, Inc.; 2017 Mar 28.

7.Wojciechowski P, Wilson K, Nazir J, et al. Efficacy and Safety of Avatrombopag in Patients with Chronic Immune Thrombocytopenia: A Systematic Literature Review and Network Meta-Analysis. Adv Ther. 2021;38(6):3113-3128. PubMed

8.Network meta-analysis to evaluate the clinical efficacy of treatments in idiopathic thrombocytopenic purpura (ITP) [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: Tavalisse (fostamatinib), 100 mg and 150 mg tablet, oral. Toronto (ON): Medison Pharma Canada Inc.; 2021 Jul 12.

9.Clinical Study Report: C-935788-049. A phase 3 open-label extension study of fostamatinib disodium in the treatment of persistent/chronic immune thrombocytopenic purpura [internal sponsor's report]. San Francisco (CA): Rigel Pharmaceuticals, Inc.; 2021 Feb 23.

10.Samson M, Fraser W, Lebowitz D. Treatments for Primary Immune Thrombocytopenia: A Review. Cureus. 2019;11(10):e5849. PubMed

11.Terrell D, Beebe L, Vesely S, Neas B, Segal J, George J. The incidence of immune thrombocytopenic purpura in children and adults: A critical review of published reports. Am J Hematol. 2010;85(3). PubMed

12.Weycker D, Hanau A, Hatfield M, et al. Primary immune thrombocytopenia in US clinical practice: incidence and healthcare burden in first 12 months following diagnosis. J Med Econ. 2020;23(2). PubMed

13.Cohen Y, Djulbegovic B, Shamai-Lubovitz O, Mozes B. The bleeding risk and natural history of idiopathic thrombocytopenic purpura in patients with persistent low platelet counts. Arch Intern Med. 2000;160(11). PubMed

14.Swinkels M, Rijkers M, Voorberg J, Vidarsson G, Leebeek FWG, Gerard Jansen AJ. Emerging Concepts in Immune Thrombocytopenia. Front Immunol. 2018;9:880. PubMed

15.Cooper N, Kruse A, Kruse C, et al. Immune thrombocytopenia (ITP) World Impact Survey (iWISh): Patient and physician perceptions of diagnosis, signs and symptoms, and treatment. Am J Hematol. 2021;96(2). PubMed

16.Nazaryan H, Liu Y, Sirotich E, Duncan J, Nazy I, Arnold D. Second-Line Therapy for Immune Thrombocytopenia: Real-World Experience in Canada. Can J Gen Intern Med. 2020;15(4):28-35.

17.Revolade (eltrombopag): 12.5 mg, 25 mg, 50 mg and 75 mg eltrombopag (as eltrombopag olamine) tablets, [product monograph]. Dorval (QC): Novartis Pharmaceuticals Canada Inc.; 2017 May 11: https://pdf.hres.ca/dpd_pm/00039266.PDF. Accessed 2021 Oct 20.

18.Nplate (romiplostim for injection): lyophilized powder for solution 250 mcg/0.5 mL and 500 mcg/1 mL [product monograph]. Mississauga (ON): Amgen Canada Inc.; 2021 Aug 04: https://pdf.hres.ca/dpd_pm/00062536.PDF. Accessed 2021 Oct 20.

19.Rituxan (rituximab): 10 mg/mL intravenous infusion [product monograph]. Mississauga (ON): Hoffmann-La Roche Ltd.; 2021 Jan 28: https://www.rochecanada.com/PMs/Rituxan/RituxanIV_PM_E.pdf. Accessed 2021 Oct 20.

20.Lucchini E, Zaja F, Bussel J. Rituximab in the treatment of immune thrombocytopenia: what is the role of this agent in 2019? Haematologica. 2019;104(6):1124-1135. PubMed

21.McGowan J, Sampson M, Salzwedel DM, Cogo E, Foerster V, Lefebvre C. PRESS Peer Review of Electronic Search Strategies: 2015 guideline statement. J Clin Epidemiol. 2016;75:40-46. PubMed

22.Grey matters: a practical tool for searching health-related grey literature. Ottawa (ON): CADTH; 2019: https://www.cadth.ca/grey-matters. Accessed 2021 Aug 8.

23.Bussel J, Arnold DM, Grossbard E, et al. Fostamatinib for the treatment of adult persistent and chronic immune thrombocytopenia: Results of two phase 3, randomized, placebo-controlled trials. Am J Hematol. 2018;93(7):921-930. PubMed

24.Drug Reimbursement Review sponsor submission: Tavalisse (fostamatinib), 100 mg and 150 mg tablet, oral [internal sponsor's package]. Toronto (ON): Medison Pharma Canada Inc.; 2021 Jul 22.

25.Ren S, Oakley JE, Stevens JW. Incorporating Genuine Prior Information about Between-Study Heterogeneity in Random Effects Pairwise and Network Meta-analyses. Med Decis Making. 2018;38(4):531-542. PubMed

26.Boccia R, Cooper N, Ghanima W, et al. Fostamatinib is an effective second-line therapy in patients with immune thrombocytopenia. Br J Haematol. 2020;190(6):933-938. PubMed

27.Cooper N, Altomare I, Thomas MR, et al. Assessment of thrombotic risk during long-term treatment of immune thrombocytopenia with fostamatinib. Ther Adv Hematol. 2021;12:20406207211010875. PubMed

28.Bussel JB, Arnold DM, Boxer MA, et al. Long-term fostamatinib treatment of adults with immune thrombocytopenia during the phase 3 clinical trial program. Am J Hematol. 2019;94(5):546-553. PubMed

29.Yang R, Lin L, Yao H, Ji O, Shen Q. Therapeutic options for adult patients with previously treated immune thrombocytopenia - a systematic review and network meta-analysis. Hematol. 2019;24(1):290-299. PubMed

30.Page LK, Psaila B, Provan D, et al. The immune thrombocytopenic purpura (ITP) bleeding score: assessment of bleeding in patients with ITP. Br J Haematol. 2007;138(2):245-248. PubMed

31.Fogarty PF, Tarantino MD, Brainsky A, Signorovitch J, Grotzinger KM. Selective validation of the WHO Bleeding Scale in patients with chronic immune thrombocytopenia. Curr Med Res Opin. 2012;28(1):79-87. PubMed

32.Signorovitch J, Brainsky A, Grotzinger KM. Validation of the FACIT-fatigue subscale, selected items from FACT-thrombocytopenia, and the SF-36v2 in patients with chronic immune thrombocytopenia. Qual Life Res. 2011;20(10):1737-1744. PubMed

33.Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting results of cancer treatment. Cancer. 1981;47(1):207-214. PubMed

34.Ware J, Kosinski M, Bjorner JB, Turner-Bowker DM, Gandek B, Maruish ME. Determining important differences in scores. User's manual for the SF-36v2 health survey. 2nd ed. Lincoln (RI): Quality Metric Inc; 2007.

35.Leung YY, Ho K, Zhu T, Tam L-S, Kun E, Li E. Testing scaling assumptions, reliability and validity of medical outcomes study short-form 36 health survey in psoriatic arthritis. Rheumatology (Oxford). 2010;49:1495-1501. PubMed

Appendix 1: Literature Search Strategy

Note that this appendix has not been copy-edited.

Clinical Literature Search

Overview

Interface: Ovid

Databases:

• MEDLINE All (1946-present)

• Embase (1974-present)

Note: Subject headings and search fields have been customized for each database. Duplicates between databases were removed in Ovid.

Date of search: August 20, 2021

Alerts: Biweekly search updates until project completion

Search filters applied: No filters were applied to limit the retrieval by study type.

Limits:

• Conference abstracts: excluded

Table 25: Syntax Guide

Multi-Database Strategy

1. (fostamatinib* or Tavalisse* or Tavlesse* or R935788 or R-935788 or R788 or R-788 or SQ8A3S5101 or 86EEZ49YVB or X9417132K8 or R 406 or R406 or R950091 or R 950091 or NSC-745942 or NSC745942 or R-935788 or R935788 or tamatinib fosdium).ti,ab,kf,ot,hw,nm,rn.

2. 1 use medall

3. *fostamatinib/ or (fostamatinib* or Tavalisse* or Tavlesse* or R935788 or R-935788 or R788 or R-788 or R 406 or R406 or R950091 or R 950091 or NSC-745942 or NSC745942 or R-935788 or R935788 or tamatinib fosdium).ti,ab,kw,dq.

4. 3 use oemezd

5. (conference review or conference abstract).pt.

6. 4 not 5

7. 2 or 6

8. remove duplicates from 7

Clinical Trials Registries

ClinicalTrials.gov

Produced by the US National Library of Medicine. Targeted search used to capture registered clinical trials.

[Search -- Studies with results | fostamatinib or Tavalisse or Tavlesse | immune thrombocytopenia OR ITP ]

WHO ICTRP

International Clinical Trials Registry Platform, produced by the World Health Organization. Targeted search used to capture registered clinical trials.

[Search terms -- fostamatinib or Tavalisse or Tavlesse | immune thrombocytopenia OR ITP]

Health Canada’s Clinical Trials Database

Produced by Health Canada. Targeted search used to capture registered clinical trials.

[Search terms -- fostamatinib or Tavalisse or Tavlesse | immune thrombocytopenia OR ITP]

EU Clinical Trials Register

European Union Clinical Trials Register, produced by the European Union. Targeted search used to capture registered clinical trials.

[Search terms -- fostamatinib or Tavalisse or Tavlesse | immune thrombocytopenia OR ITP]

Grey Literature

Search dates: August 10-16, 2021

Keywords: fostamatinib, Tavalisse, Tavlesse, ITP, immune thrombocytopenia

Limits: none

Updated: Search updated prior to the meeting of the CADTH Canadian Drug Expert Committee (CDEC)

Relevant websites from the following sections of the CADTH grey literature checklist Grey Matters: A Practical Tool for Searching Health-Related Grey Literature were searched:

• Health Technology Assessment Agencies

• Health Economics

• Clinical Practice Guidelines

• Drug and Device Regulatory Approvals

• Advisories and Warnings

• Drug Class Reviews

• Clinical Trials Registries

• Databases (free)

• Health Statistics

• Internet Search

• Open Access Journals.

Appendix 2: Excluded Studies

Note that this appendix has not been copy-edited.

Table 26: Excluded Studies

Appendix 3: Description and Appraisal of Outcome Measures

Note that this appendix has not been copy-edited.

Aim

To describe the following outcome measures and review their measurement properties (validity, reliability, responsiveness to change, and MID):

• ITP bleeding scale (IBLS)

• World Health Organization (WHO) bleeding scale

• Short Form (36) Health Survey Version 2 (SF-36v2)

Findings

Table 27: Summary of Outcome Measures and Their Measurement Properties

GYN = gynecological; IQRs = The interquartile ranges; MID = minimal important difference.

ITP Bleeding Scale

The IBLS is an assessment method for objective quantification of bleeding symptoms among ITP patients. IBLS is comprised of 11 anatomical site-specific bleeding grades from 0 (none) to 2 (marked bleeding) and is assessed by history over the previous week (Hx) for all, and by physical examination (PE) for 2 of these sites. The sites are – skin (PE), oral (PE), skin (Hx), oral (Hx), epistaxis, gastrointestinal (GI), urinary (U), gynecological (GYN), pulmonary, intracranial hemorrhage, subconjunctival hemorrhage.

In a pilot study conducted between 2004 to 2005 among 65 ITP patients, IBLS had been used to analyze the correlation of platelet variables with bleeding. In this study, 92% of the sites were graded identically, demonstrating a good inter-observer reliability. The values for Kappa statistics were – 0.71 for skin (Hx), 0.66 for skin (PE), 0.52 for oral (Hx), 0.46 for oral (PE), 0.58 for epistaxis, and 0.78 for GYN.³⁰ In addition to this study, IBLS had been evaluated in 2 long-term, phase III clinical trials named RAISE and EXTEND, where participants were chronic ITP patients taking eltrombopag. In these studies, the clinical investigator assessed the blooding grade based on verbal responses and physical examination.³¹

Intraclass correlation coefficients for test–retest reliability were calculated using 2 consecutive bleeding grades corresponding to the nearest platelet counts for each patient. The interquartile ranges (IQRs) were from - 0.8% to + 1.3% with ICCs 0.63, and from - 4.7% to + 3.4% with ICCs 0.77 for the IBLS in the RAISE and EXTEND studies, respectively.

Construct validity was assessed by determining inter-instrument correlations (item-to-item and item-to-domain correlations) between the WHO bleeding scale and the IBLS. The relationship between the WHO bleeding scale and the IBLS had been described by comparing mean WHO grades with mean IBLS summary scores at baseline, as well as by assessing the associations between scores on each patient-reported outcome instrument and platelet counts at baseline and last on-treatment. In addition, WHO grades were cross-classified with item-level responses for the IBLS items to demonstrate the relationship between the 2 scales. A positive correlation was observed between the WHO bleeding scale and the IBLS, showing a similar association between platelet counts and severity of bleeding. However, the IBLS demonstrated more capability in capturing more details about bleeding than the WHO bleeding scale.

The responsiveness was assessed by calculating the differences in grades from baseline to last-on-treatment evaluation among patients with a platelet count response. The responsiveness among patients were computed using 3 indices: effect size = D/SD⁰ , standardized response mean = D/ SD*, and responsiveness statistic = D/SD^# . Here D denotes the mean score change of interest (i.e., mean change from baseline among patients with platelet count response), SD⁰ denotes the SD of scores at baseline, SD* is the SD of D, and SD^# is the SD of D among patients with no response to treatment. However, the responsiveness measures assessed in the study demonstrated a statistical property of the bleeding scales in the population under study and could not be used as a description of how patients responded to therapy in the RAISE and EXTEND studies. In the RAISE study containing 129 patients, the responsiveness scores comprising the effect size, standardized response and responsiveness statistic were 0.719, 0.741 and 0.640, respectively, whereas in the EXTEND study among 71 patients, the effect size, standardized response and responsiveness statistic were 0.560, 0.561 and 0.506, respectively, showcasing moderate responsiveness for the IBLS in both studies.³¹

The MID was not assessed for IBLS among ITP patients in any study.

WHO Bleeding Scale

The WHO bleeding scale is an instrument used to classify bleeding among ITP patients on a 5-point scale. The scale grading follows as: 0 (no bleeding), 1 (petechiae), 2 (mild blood loss), 3 (gross blood loss), and 4 (debilitating blood loss).³¹ Originally developed for bleeding assessment among cancer patients,³³ the performance of WHO bleeding scale had been evaluated among chronic ITP patients taking eltrombopag in 2 long-term, phase III clinical trials, RAISE and EXTEND studies.³¹

Intraclass correlation coefficient for test–retest reliability were calculated using 2 consecutive bleeding grades corresponding to the nearest platelet counts for each patient. The interquartile ranges (IQRs) were from - 0.8% to + 1.3% with ICCs 0.75, and from - 4.7% to + 3.4% with ICCs 0.70 for the WHO bleeding scale in the RAISE and EXTEND studies, respectively.

Construct validity was assessed in 2 ways - by determining inter-instrument correlations (item-to-item and item-to-domain correlations) between the WHO bleeding scale and the IBLS and between the WHO bleeding scale and platelet counts, and by describing the relationship between the scale and clinical outcomes. A positive correlation was observed between the WHO bleeding scale and IBLS, showing a similar association between platelet counts and severity of bleeding while assessing inter-instrument correlations. Moreover, during known group comparison assessment, significant associations (p<0.05) were observed between the WHO bleeding scale and many clinical outcomes.

The responsiveness was assessed by calculating the differences in grades from baseline to last-on-treatment evaluation among patients with a platelet count response. The responsiveness among patients were computed using 3 indices: effect size = D/SD⁰, standardized response mean = D/ SD*, and responsiveness statistic = D/SD^#. Here D denotes the mean score change of interest (i.e., mean change from baseline among patients with platelet count response), SD⁰ denotes the SD of scores at baseline, SD* is the SD of D, and SD^# is the SD of D among patients with no response to treatment. However, the responsiveness measures assessed in the study demonstrated a statistical property of the bleeding scales in the population under study and could not be used as a description of how patients responded to therapy in the RAISE and EXTEND studies. In the RAISE study containing 129 patients, the effect size, standardized response and responsiveness statistic were 0.714, 0.745 and 0.560, respectively, indicating moderate responsiveness. In the EXTEND study among 71 patients, the effect size, standardized response and responsiveness statistic were 0.622, 0.487 and 0.588, respectively, showcasing moderate responsiveness for 2 responsiveness indices, and just below the 0.50 threshold for moderate responsiveness for 1 index.

The estimated MID for the WHO bleeding scale ranged from 0.33 to 0.40.³¹

Short Form (36) Health Survey

SF-36 is a generic health assessment questionnaire that has been used in clinical trials to study the impact of chronic disease on HRQoL. SF-36 consists of 8 domains: physical functioning, role physical, bodily pain, general health, vitality, social functioning, role emotional, and mental health. SF-36 also provides 2 component summaries: the physical component summary (SF-36 PCS) and the mental component summary (SF-36 MCS), which are created by aggregating the 8 domains. The SF-36 PCS, SF-36 MCS and 8 domains are each measured on a scale of 0 to 100, with an increase in score indicating improvement in health status. In general use of SF-36, a change of 2 to 4 points in each domain or 2 to 3 points in each component summary indicates a clinically meaningful improvement as determined by the patient.³⁴ The summary scales are scored using norm-based methods, with regression weights and constants derived from the general US population. Both the PCS and MCS scales are transformed to have a mean of 50 and an SD of 10 in the general US population. Therefore, all scores above/below 50 are considered above/below average for the general US population.³⁵

The validity, reliability, and responsiveness of SF-36v2 had been assessed in 2 clinical trials, RAISE and EXTEND studies, prescribing eltrombopag to patients previously treated for chronic ITP. RAISE was a 6-month, phase III, randomized, double-blind, placebo-controlled study with 197 ITP patients, whereas EXTEND was an open-label extension study containing 154 patients. In the RAISE study, SF-36v2 PCS and MCS mean scores were below but within 1 SD of the US population standardized mean, as well as the mean scores for the 7 out of 8 domains. Sufficient and acceptable internal consistency had been reported, demonstrating all SF-36 item-to-domain score correlations >0.20, and Cronbach alpha values for SF-36v2 domains ≥0.75, for all items and if each item was deleted from scale, in both RAISE and EXTEND studies. More specifically, Cronbach alpha values for SF-36v2 were between 0.75 and 0.94 at baseline and between 0.83 and 0.95 at the last assessment in RAISE, and between 0.78 and 0.94 at baseline and between 0.79 and 0.96 at the last assessment in EXTEND.

ICCs for test–retest reliability evaluation in clinically stable patients were >0.7 in both RAISE and EXTEND studies. In RAISE, this value was applicable for physical function, general health, and vitality domains (n = 50–55), and in EXTEND for all domains of SF-36, except bodily pain and emotional role (n = 126–132). Here the calculation of ICC for clinically stable patients were based on assessments from 2 consecutive pair of visits when platelet counts for each patient were considered most similar (mean of 42 days for RAISE and 39–43 days in EXTEND). However, no absolute degree of similarity had been imposed between studied pair of visits. During sensitivity analyses, ICCs were calculated using a subgroup of patients with ≤15% change in platelet counts between 2 consecutive visits (mean of 49–52 days for RAISE and 45–50 days in EXTEND). For sensitivity analysis, ICCs in clinically stable patients were ≥0.72 for all domains and summary measures of SF-36-v2, except social function and emotional role.

Construct validity of SF-36v2 was assessed by testing hypotheses about relationships with other instruments and with clinical outcomes, and was supported by moderate to strong score correlations between scores at baseline, and between the change scores of the PRO measures in both studies. While evaluating the longitudinal construct validity of measures by stratifying patients into responders or nonresponders and comparing the change score on each measure between groups based on magnitude of effect, a statistically significant difference between responders and nonresponders was observed for SF-36v2. Regarding the responsiveness, SF-36v2 had been reported to be less responsive compared to the disease-specific measures of fatigue based on the ability to capture change.³²

No MID had been assessed for SF-36v2 among ITP patients.

Pharmacoeconomic Review

Executive Summary

The executive summary comprises 2 tables (Table 1 and Table 2) and a conclusion.

Table 1: Submitted for Review

ITP = immune thrombocytopenia; NOC = Notice of Compliance; TPO-RA = thrombopoietin receptor agonist.

Table 2: Summary of Economic Evaluation

ICER = incremental cost-effectiveness ratio; ITP = immune thrombocytopenia; LY = life-year; NMA = network meta-analysis; QALY = quality-adjusted life-year; TPO-RA = thrombopoietin receptor agonist.

Conclusions

The CADTH clinical review found fostamatinib led to a modest improvement in platelet count response compared to placebo among patients with chronic immune thrombocytopenia (ITP). Fostamatinib was generally well tolerated compared to placebo, but its effect on clinical outcomes important to patients such as bleeding rates, symptoms, and quality of life remains unclear. The sponsor-submitted indirect treatment comparison found fostamatinib was favoured over rituximab in achieving overall platelet response, but this study had a number of limitations that affected internal and external validity. Additionally, the comparative efficacy and safety of fostamatinib to other active ITP therapies, such as thrombopoietin receptor agonists (TPO-RAs), was not explored. Given the absence of evidence, the comparative efficacy and safety of fostamatinib versus other relevant ITP treatments, and effect on clinical outcomes important to patients, remain unclear.

CADTH undertook reanalyses to address limitations with the sponsor’s submission. These reanalyses included: assuming that the extrapolation of transition probabilities is based on the number of individuals at risk; altering the rate of rescue medication usage; and revising blood platelet health state utilities. Additionally, the reanalyses corrected a number of coding and modelling errors, primarily related to the way that probabilities were translated across difference cycle lengths (e.g., weekly probabilities into 4-week cycles).

CADTH was unable to estimate an ICER for the Health Canada indication for fostamatinib, as the economic model did not reflect this indication and was missing relevant comparators (i.e., TPO-RAs, long-term steroids, and immunosuppressant drugs). For the sponsor’s reimbursement request, the CADTH base-case sequential analysis found fostamatinib was more effective and more costly than watch and rescue (incremental quality-adjusted life-years [QALYs]: 0.77, incremental cost: $164,368), and had an incremental cost-effectiveness ratio (ICER) of $212,783 per QALY. Compared to watch and rescue, rituximab was dominated (i.e., had higher costs and lower effectiveness). The probability that fostamatinib was cost-effective at a $50,000 willingness-to-pay threshold was less than 0.01%. Probabilistic results were hampered by how the sponsor propagated parameter uncertainty throughout the model. A price reduction of 60.2% for fostamatinib is needed to achieve an ICER of $50,000 per QALY in the requested reimbursement population.

The 2 main drivers of the model findings include the acquisition costs of fostamatinib, and the risk of rescue events and subsequent need for rescue therapy in both the fostamatinib and watch-and-rescue strategies. The latter risk is informed by limited data from the FIT trials, increasing the uncertainty in the economic findings.

Stakeholder Input Relevant to the Economic Review

This section is a summary of feedback received from the patient groups, registered clinicians, and drug plans that participated in the CADTH review process.

CADTH received 1 patient input submission from the Platelet Disorder Support Association (PDSA) for this review. Patient comments from 2018 to present were collected from the PDSA Facebook page and appear to include primarily international patients. Patients comments included in the submission noted increasing platelet counts while using fostamatinib, reduced steroid use, manageable side effects such as elevated blood pressure and chronic diarrhea, and a better individual response than the patients had experienced on previous therapies, which included steroids, IV immunoglobulin (IVIG), rituximab, a splenectomy, and TPO-RAs. The PDSA noted that ITP affects the overall health-related quality of life (HRQoL) of patients and their families, with a constant risk of life-threatening bleeding, elevated levels of fatigue, anxiety, depression, physical pain, sleep disturbances, and feelings of isolation and inadequacy due to activity restrictions. The PDSA also noted that patients often do not have a choice of treatment as they may not respond well to some therapies and may be unable to afford others, that fostamatinib is a daily pill that is easier to use than injections or infusions requiring travel to a clinic or hospital, and that an adequate preventive therapy might avoid expenses associated with ITP such as potential hospitalizations, life-support, treatments such as IVIG, and the long-term effects of steroid use.

One joint clinician input submission, commissioned by Accelera Canada in partnership with Advocacy Solutions, was received. Nineteen Canadian hematologists participated in the submission, which noted that the treatment paradigm for ITP is not uniform across Canada, due to differential access to therapies, and that most treatments are not approved by Health Canada but are required by public plans to access Health Canada–approved treatments. The clinician submission indicated fostamatinib should be used as a second or subsequent line of therapy after corticosteroids and before a splenectomy, immunosuppressant drugs, or rituximab, and be comparable to maintenance treatments such as TPO-RAs. Patients are expected to respond better to fostamatinib earlier in their treatment course.

Drug plan input noted that, in some jurisdictions, a patient must have previously used rituximab to access a TPO-RA; the sponsor’s reimbursement request has placed rituximab after TPO-RAs in its proposed treatment paradigm. The plans also noted that evidence for fostamatinib appears to have similar issues to those of eltrombopag, which was not recommended by the CADTH Common Drug Expert Committee due to concerns with the primary outcome of its trials, a lack of comparative evidence to other available treatments, and a lack of cost-effectiveness. Additionally, the plans expressed uncertainty about the assumed market uptake of fostamatinib in the budget impact analysis (BIA), as well as the cost of rituximab given the availability of biosimilar products.

Two of these concerns were addressed in the sponsor’s model:

• Based on data from the FIT trials, differential adverse-event rates are accounted for.

• HRQoL is captured in the model.

In addition, CADTH conducted scenario analyses in the BIA exploring uncertainty in the proportion of patients who would qualify for fostamatinib, the expected market uptake of fostamatinib, and the price of rituximab and blood products.

CADTH was unable to address the following concerns raised from stakeholder input:

• The treatment order assumed by the sponsor, in which fostamatinib is used after TPO-RAs or in place of TPO-RAs in jurisdictions that do not fund them, rather than in an earlier line of therapy.

• The lack of direct clinical evidence comparing fostamatinib to active treatments in use for chronic ITP in Canada.

• As multiple comparators were missing, the cost-effectiveness of fostamatinib as a second-line therapy could not be evaluated.

Economic Review

The current review is for fostamatinib (Tavalisse) for adult patients with chronic ITP (> 12 months), who are either resistant or refractory to previous lines of treatment.

Economic Evaluation

Summary of Sponsor’s Economic Evaluation

Overview

The sponsor submitted a cost-utility analysis of fostamatinib compared to rituximab and a treatment of watch and rescue in adult patients with chronic ITP (> 12 months), who are either resistant or refractory to previous lines of treatment.¹ The modelled population resembled the Canadian ITP population and matched the reimbursement request, but does not align with the Health Canada indication as several comparators were missing.² Two subgroup analyses were conducted depending on whether a patient had received treatment with TPO-RAs. A cost-utility analysis was taken from the perspective of the Canadian publicly funded health care system.

Fostamatinib is available as 100 mg or 150 mg tablets in bottles of 60 tablets. The recommended dosage of fostamatinib is 100 mg twice daily,² and if after 4 weeks platelet counts do not increase to at least 50,000/μL the dosage is increased to 150 mg twice daily. As treatment is continued, it should target the lowest dose of fostamatinib required to achieve a platelet count of 50,000/μL. According to clinical experts consulted by CADTH, the majority of patients receive a dosage of rituximab at 375 mg/m² administered weekly for 4 weeks, while a minority received 100 mg administered weekly for 4 weeks. Watch and rescue consisted of monitoring patients who are not receiving therapy and administering rescue medication when platelet counts drop below unsafe thresholds, or during bleed events. Rescue medication could consist of IVIG, IV methylprednisolone, platelet transfusion, oral dexamethasone, or oral prednisolone.

The analysis assumed no administration costs for fostamatinib as it was administered orally. Administration costs for rituximab consisted of inpatient IV infusion costs. Administration costs were considered for rescue medication that was administered via infusion. The total drug acquisition costs per 28-day model cycle for fostamatinib was $5,600 for those receiving 200 mg daily and $8,400 for those receiving 300 mg daily. The total drug acquisition costs for rituximab during the first 28-day model cycle was $9,504 for those receiving doses of 375 mg/m² and $1,188 for those receiving 100 mg. There were no drug acquisition costs for rituximab beyond the first cycle. The drug acquisition costs for rescue therapies included $4,657 for IVIG, $41 for methylprednisolone, $101 for platelet transfusion, $7 for oral dexamethasone, and $2 for oral prednisolone. The probability of receiving each specific rescue medication depended on platelet counts and the treatment received, with individuals being able to receive more than 1 drug per rescue event. Wastage was considered in the economic evaluation.

The clinical outcomes modelled included QALYs, life-years, incidence of severe disability from intracranial hemorrhage (ICH), and adverse events. The economic evaluation used a discount rate of 1.5% per year for both costs and health outcomes.

Model Structure

A cohort state transition (Markov) model was developed with health states that capture the chronic nature of treatment-resistant ITP. Most health states were defined by the surrogate outcome of platelet count, which was linked in the model to the frequency of bleeding events and the need for rescue medication. The model aims to capture the clinical benefit through an increase in platelet count, which arises from effective disease control and results in reduced morbidity and mortality. The health states of the model therefore capture a patient’s differing levels of platelet counts, as well as the long-term sequelae resulting from severe bleed events. The model has 4 mutually exclusive health states: “nonresponse” (platelet count < 30,000/µL of blood), “partial response” (platelet count between 30,000/µL and 50,000/µL of blood), “response” (platelet count > 50,000/µL of blood) and “death”; patients can transition between any of the non-dead states. All patients who enter the model are assumed to be in the nonresponse health state. The full model structure with all possible transitions is presented in Figure 1.

The model has several key assumptions. In the model, patients could have a lack of response to treatment, which is defined as having a blood platelet count of less than 30,000/µL. Patients who respond to treatment transition into either the response or partial response health states; transition probabilities are informed by the FIT clinical trials and the results of a network meta-analysis (NMA). The model captures the number of bleed events (outpatient bleeds and severe bleeds resulting in inpatient care) and rescue events within each health state to determine the differences in costs, mortality, and quality of life. The model assumes the 28-day cycle length is sufficiently short if it is less than the shortest period in which 2 bleed events could realistically occur, while still being sufficiently long to capture the length of rescue treatment. A patient can therefore only have 1 bleed event per cycle and 1 rescue event. The model assumes all patients have a platelet count–dependent risk of an ICH and patients who experience an ICH will consequently have a severe disability (a modified Rankin scale score of 4 or 5) for the remaining cycles. The model assumes all patients have a risk of death, including a risk of death from severe bleed events and non–disease-specific factors. A half-cycle correction is applied to both costs and health outcomes to account for the fact that events do not occur precisely at the beginning or the end of each cycle, and instead could occur at any point during the cycle.

Model Inputs

The model structure, clinical parameters, and model assumptions were informed by the FIT1,³ FIT2,⁴ and FIT3⁵ clinical trials. The FIT1 and FIT2 trials were 24-week phase III multi-centre, randomized, double-blind placebo studies in which participants were assigned to treatment with fostamatinib or a placebo watch-and-rescue group. The fostamatinib group was given 100 mg twice a day, with the option to increase to 150 mg twice a day after 4 weeks, depending on platelet count. In this economic analysis, the sponsor used efficacy data for the placebo group of the FIT trials to inform the first 4 weeks of the watch-and-rescue strategy. As data from the FIT1 and FIT2 trials were only available for up to 24 weeks, further efficacy data from the FIT3 trial was used up to 60 months. The FIT3 trial was a phase III, multi-centre, single-arm, open-label study clinical trial examining the long-term effects of fostamatinib. Individuals enrolled who had platelet counts of at least 50,000/μL started treatment with the same dosage and regimen as the randomized studies, or 100 mg twice a day for individuals who were nonresponders in the FIT1 and FIT2 trials. As the FIT trials were conducted worldwide, the baseline patient characteristics were sourced from Canadian-specific values for generalizability to the target population. The population in the economic model was 57% female and 43% male, and patients entered the model at an average age of 41, based on the mean age of patients with ITP observed in the McMaster ITP registry (2020).⁶ Canadian population-level mortality estimates were inflated using disease-specific mortality hazard ratios accounting for post-stroke disability, hemorrhaging, and infection. Hazard ratios were dependent on platelet counts.

Health-related quality of life was sourced from the literature and assigned based on platelet counts and severe disability. In the reference case, the economic model also assigned a disutility for caregivers of patients who experienced severe disability after an ICH. Utility values sourced from the literature were adjusted to the baseline age of 41 years using methods outlined in Ara and Brazier.⁷ Utility values associated with platelet counts were sourced from Szende et al.,⁸ a UK study that estimated health-utility values associated with ITP using time trade-off (TTO) methods. Disutility due to disability after ICH for patients and caregivers was sourced from Dewilde et al.,⁹ a study in Belgium that used the EQ-5D 3-Levels questionnaire to elicit health-utility values. Health-disutility values for severe bleed events and adverse events were sourced from published literature, which used a mix of TTO, EQ-5D, and standard gamble elicitation methods.^7,10-14 Health-disutility values for severe bleed events were calculated by subtracting the baseline utility for ICH from the utility post–severe bleed event. The rate of adverse events for patients receiving fostamatinib or watch and rescue were sourced from the FIT trials.¹⁵ The rate of adverse events for patients receiving rituximab was sourced from Ghanima et al.¹⁶

The model considered the following cost components: drug treatment cost, costs of rescue treatment, cost of pre-surgical prophylaxis, health-state costs based on clinical events and routine resource use, and adverse-event costs. All costs assumed a 2020 cost year with costs sourced from before 2020 inflated using the consumer price index.¹⁷ Drug acquisition costs were sourced from provincial formularies and health care databases, and calculated using Canadian dosing regimens.^18,19 The cost of wastage for rituximab was incorporated, and dosing was based on the body surface area of the average member of the Canadian population. Administrative costs for rituximab were sourced from Tam et al.²⁰ and based on the average length of infusions.²¹ Fostamatinib was assumed to have no administrative costs as it was administered orally. Rescue medication dosing was sourced from guidelines and the literature,^22,23 while unit costs were obtained from the literature and provincial formularies.^19,24,25 Administrative costs for rescue medications were considered. The rates of rescue medication in the FIT trials were used to inform usage rates in the economic analysis. Bleed-event rates were obtained from the literature and costs were sourced from schedules of benefits, administrative databases, and the literature.^18,26-28 For ICH, routine resource use rates were obtained from key opinion leaders, varying both by platelet count and model cycle. Resource unit costs for hematologist consultations, blood tests, and biochemistry were sourced from national cost databases. Costs associated with severe disability due to ICH were obtained from Goeree et al.²⁹ Other adverse-event costs were sourced from Bussel et al.¹⁵ and Ghanima et al.¹⁶

An NMA was conducted to provide an indirect treatment comparison between fostamatinib and rituximab through a common comparator of placebo (watch and rescue). The results of the NMA are summarized and presented as odds ratios (ORs). Specifically, the ORs of rituximab (375 mg/m²) versus fostamatinib, rituximab (100 mg) versus fostamatinib, and placebo versus fostamatinib are generated by the NMA. The observed transition probabilities for patients receiving fostamatinib in the FIT trials are used as the reference transitions and then the ORs from the NMA are applied to these probabilities for both rituximab and watch and rescue. Specifically, the ORs are applied to the probabilities of achieving either a partial or full platelet response, as ORs computed within the NMA are based on a definition of response of the achievement of a platelet count of at least 30,000/μL. The probability of nonresponse is calculated as 1 minus the probabilities of achieving either a partial or full platelet response. The transition matrices for both rituximab and watch and rescue, computed using the ORs provided by the NMA, are applied in the appropriate cycles in the Markov trace. The model assumes that these ORs are time-invariant and are applied in the same fashion each cycle.

Summary of Sponsor’s Economic Evaluation Results

Base-Case Results

The sponsor’s base-case analysis was run probabilistically for 5,000 iterations. Fostamatinib was found to have the highest expected cost ($1,084,675) and the highest QALY gain (14.977). Rituximab was found to be more costly and less effective than watch and rescue and was therefore dominated. Compared to watch and rescue, the ICER for fostamatinib is $149,029 per QALY gained. The probability that fostamatinib was cost-effective at a willingness-to-pay threshold of $50,000 per QALY is 1.08%.

Sensitivity and Scenario Analysis Results

Seven scenario analyses were conducted to evaluate the impact of adjusting key parameter values or assumptions. These scenarios modified the discount rates, reduced the time horizon, used inputs from the FIT trials stratified by whether patients had received TPO-RA therapy, and varied how the NMA results were implemented and the source for the rate of the bleed events. When restricting the FIT trials to patients who had not previously received TPO-RA therapy, the ICER for fostamatinib compared to watch and rescue was $133,171 per QALY gained. When restricting the FIT trials to patients who had received TPO-RA therapy, the ICER was $145,978 per QALY gained when comparing fostamatinib to watch and rescue. In all scenarios, except for varying the implementation of NMA results, rituximab was dominated by watch and rescue.

CADTH Appraisal of the Sponsor’s Economic Evaluation

CADTH identified several key limitations to the sponsor’s analysis that have notable implications on the economic analysis:

• Missing comparators: Based on feedback from clinical experts consulted for this review, relevant comparators were excluded by the sponsor, specifically TPO-RAs, long-term steroids, and immunosuppressant agents. Additionally, the clinical experts consulted by CADTH did not agree that watch and rescue was an appropriate therapy at this late stage of therapy, and would instead prescribe an immunosuppressant drug the patient had not previously tried.

  • CADTH was unable to address this limitation in reanalyses. As such, the cost-effectiveness of fostamatinib relative to these other comparators is unknown.

• Unknown cost-effectiveness for the Health Canada indication: The modelled population aligned with the reimbursement request of adult patients with ITP who had an insufficient response to a TPO-RA in jurisdictions where TPO-RA reimbursement is available, or after failure of corticosteroids and other earlier-line treatments in jurisdictions where TPO-RA reimbursement is not available. This does not reflect the Health Canada indication of adult patients with ITP who have had insufficient response to previous treatments, as later lines of therapy beyond first-line corticosteroids use vary; patients could therefore have an insufficient response to therapies other than TPO-RAs, making various treatments relevant comparators. To evaluate the full indication, additional comparators need to be included, and the cost-effectiveness of fostamatinib for the Health Canada indication is unknown.

  Additionally, input received from clinicians indicated that the preferred place in therapy for fostamatinib would be as early as the second line. Should fostamatinib be reimbursed as second-line therapy, the appropriate comparators would include rituximab, TPO-RAs, immunosuppressant drugs, and a splenectomy.

  • CADTH was unable to address this limitation in its reanalysis.

• Uncertainty when extrapolating the duration of treatment response: Due to the limited follow-up of the FIT1 and FIT2 trials for treatment response, the probability of remaining in a state with a platelet count of greater than 50,000/μL after 24 weeks in the model was estimated using data from 13 responders in the FIT3 trial. An exponential distribution was fit to the maximum follow-up time (1,661 days) as less than 50% of the 13 responders had experienced an event. This model parameter was fixed in the sponsor’s submission, and alternative distribution fits were not provided.

  • CADTH was able to partially address this limitation. CADTH conducted a scenario analysis using the sponsor’s exponential distribution fit to the mid-point between the minimum and maximum follow-up, which was 721 days. CADTH also conducted a scenario analysis by fitting an exponential distribution to the observed values for overall responders in FIT3 at the end of follow-up (65% of responders remained responders at 28 months). CADTH also conducted a scenario analysis by fitting an exponential distribution on the 24-month percentage of the 13 responders still responding (77% at 24 months for blood platelet counts > 30,000/μL). Given the lack of alternative parametric models, CADTH was unable to fully address the uncertainty surrounding the duration of treatment response for responders.

• Uncertainty when extrapolating platelet response data: Beyond the clinical trial follow-up period (24 weeks), the probability of transitioning from a platelet count of between 30,000/μL and 50,000/μL to fewer than 30,000/μL was based on a weighted average of the transition probabilities for weeks 1 to 4, weeks 5 to 12, and weeks 13 to 24. It is unclear if this is appropriate, as the sponsor modelled a time effect and the 3 time points had unique transition probabilities. Additionally, the sponsor’s weighted average assumed that 60% of the data available to inform transition probabilities beyond week 24 was sourced from only 20 of the 102 patients in the fostamatinib arm of the FIT1 and FIT2 trials. Of those 20 patients, only 5 experienced a transition to a different blood platelet count state during that period.

  • To address this limitation, CADTH modified the base case by informing the transition probabilities beyond the trial follow-up period (24 weeks) using data from the transitions observed from week 1 to week 24 weighted by the number of days at risk. While this approach does not consider a time treatment effect for fostamatinib, it ensures that more patient data from the FIT trials are used to inform transition probabilities. Additionally, CADTH conducted a scenario analysis in which the transition probabilities for model cycle 6 and onward are equal to those in model cycles 4 to 5.

• Utilities incorrectly sourced from a variety of methods: The sponsor’s economic model uses utility estimates from the literature that were derived from a variety of elicitation techniques. The primary utility estimate of HRQoL associated with platelet count was sourced using an unvalidated TTO method. CADTH guidelines recommend the use of an indirect method, such as the EQ-5D, healthy utility index, or Short Form Six-Dimensions Health Utility Survey, based on a generic classification system.³⁰ The disutility due to disability after ICH, severe bleeds, and adverse events were elicited from TTO, ED-5D, and standard gamble elicitation methods. The sponsor did not attempt to crosswalk its values into a single measure. Given this limitation, the utility estimates, and consequently the QALY estimates, are likely to be imprecise and unreliable.

  • CADTH sourced utility estimates from a study that relied on an indirect elicitation method (EQ-5D) to estimate health utilities across a range of platelet counts among patients with ITP.³¹ The CADTH base case was modified to include these utilities, which were derived using a weighted average of the reported platelet counts to represent the model’s platelet count health states. The utilities used were 0.81, 0.82, and 0.815 for the health state associated with blood platelet counts of fewer than 30,000/μL, 30,000/μL to 50,000/ μL, and greater than 50,000/μL, respectively.

• Inappropriate use of a constant benefit for fostamatinib compared to watch and rescue and rituximab over the lifetime model time horizon: Clinical experts consulted by CADTH indicated that the response rate for fostamatinib is unlikely to be constant over time (i.e., likely to decrease). The sponsor’s model assumes that the relative benefit of fostamatinib over watch and rescue will be the same at year 10 as it is in year 1. This assumption is likely generating a lower ICER for fostamatinib compared to watch and rescue.

  • CADTH explored this limitation by conducting 2 scenario analyses that limited the time horizon of the cost-effectiveness model to 5 years (the duration of the FIT3 trial) and 10 years.

• Possibly overestimated rate of rescue medication use: In the sponsor’s submission, the rate of rescue medication among responders (blood platelet counts > 50,000/μL) treated with watch and rescue is 7 times higher than for those who are receiving rituximab or fostamatinib (0.603 events per model cycle compared to 0.072). This value was estimated from a small subgroup of individuals from the FIT1 and FIT2 trials. Clinicians consulted by CADTH indicated that this value was unlikely to represent typical clinical practice.

  • To address this limitation, CADTH modified the base-case analysis so that the rate of rescue medication is conditional on response status as opposed to response status and treatment. Additionally, CADTH created a scenario analysis in which the risk of rescue medication in the watch-and-rescue group is equal to 1.70 times the risk in the rituximab or fostamatinib treatments groups. This number represents the overall increase in the rate of rescue medication usage from those who received placebo compared to fostamatinib in the FIT1 and FIT2 trials.

• Limitations associated with indirect evidence: The NMA used by the sponsor did not differentiate between levels of platelet response. Instead, an outcome of overall platelet response was used, and the sponsor’s NMA did not estimate a platelet count–specific OR. Further, the sponsor applied its pooled estimate so that the relative benefit of fostamatinib was the same for all blood platelet count states. This is likely to generate more optimistic estimates of fostamatinib as the largest treatment benefit is seen in partial platelet response, and by lumping the treatment effects of partial and full responses, the treatment effect of full response may be artificially inflated.

  • CADTH was unable to fully address this limitation. CADTH conducted a scenario analysis using a different analysis from the NMA (i.e., analysis 3) that generated more conservative estimates of the benefit of fostamatinib.

• Failure of parameter uncertainty to accurately reflect uncertainty around the ICER: Incorporation of parameter uncertainty did not follow CADTH guidelines, as the sponsor did not source uncertainty estimates for most of the parameters. Instead, 98% of model parameters used an arbitrary standard error of 20% of the mean. This is also the case for model inputs for which the sponsor had generated standard errors, such as the OR from the NMA the sponsor conducted. Additionally, uncertainty in the remaining key model parameters were improperly implemented, which reduced uncertainty in the modelled results. For example, the uncertainty surrounding the transition probabilities is modelled using a Dirichlet distribution, with the input being the hypothetical modelled cohort (e.g., 1,000 patients) multiplied by the transition probability. As the hypothetical population increases, the uncertainty is reduced in the transition matrices. Given this uncertainty, it is unclear whether cost-effectiveness outcomes are underestimated or overestimated; however, improper incorporation of uncertainty biases the cost-effectiveness outcomes.

  • CADTH was unable to address this limitation in its reanalysis.

• Use of surrogate outcomes: The health states in the model obtained from the FIT trials are defined by blood platelet counts, which are surrogate outcomes for survival and HRQoL. While the clinical experts consulted by CADTH indicated that these were appropriate proxies for effective disease control, this introduces additional uncertainty to the sponsor’s model in how effective disease control translates to estimates of life-years and QALYs.

  • CADTH was unable to address this limitation in its reanalysis.

Additionally, the following key assumptions were made by the sponsor and have been appraised by CADTH (Table 4).

Table 4: Key Assumptions of the Submitted Economic Evaluation Not Noted as Limitations to the Submission

NMA = network meta-analysis.

CADTH Reanalyses of the Economic Evaluation

Base-Case Results

The CADTH base case was derived by making changes in model parameter values and assumptions, in consultation with clinical experts (Table 5).

CADTH’s base-case results are presented in Table 7 and stepped reanalysis in Table 6. Disaggregated results of the CADTH reanalysis are presented in Table 12. In CADTH’s base case, fostamatinib was associated with the highest total discounted costs ($924,745) and QALYs (15.32) over the lifetime horizon. The ICER comparing fostamatinib to watch and rescue was $212,783 per QALY. Rituximab was dominated (i.e., had higher costs and fewer QALYs) than the watch-and-rescue strategy. The probability that fostamatinib was cost-effective at a willingness-to-pay threshold of $50,000 per QALY was less than 0.01%. The percentages of QALYs generated within the 5-year trial period for FIT3 were 24.1%, 23.8%, and 24.0% for the watch-and-rescue, fostamatinib, and rituximab strategies, respectively.