Fedratinib for the treatment of myelofibrosis: a critical appraisal of clinical trial and “real-world” data

Fedratinib is an oral small-molecule kinase inhibitor that is approved for the treatment of primary and post-polycythemia vera (PV) and post-essential thrombocythemia (ET) myelofibrosis (MF) across the globe. While primarily a JAK2 inhibitor with a 50% inhibitory concentration (IC₅₀) of only 3 nM, fedratinib also targets other JAK proteins namely JAK1, JAK3, and TYK2, albeit with 35–334 fold less potency [1]. Importantly, fedratinib displays off-target effects inhibiting FLT3, a tyrosine kinase expressed on hematopoietic stem cells as well as BRD4, a BET family member with epigenetic regulatory function and which blocks NF-kB hyperactivation and inflammatory cytokine production [2]. The combined inhibitory effects of the drug retarded cell proliferation and induced apoptotic cell death leading to increased survival in murine models of jak2 V617F mutation-induced myeloproliferation [3]. These observations led to a clinical drug development program, and ultimately regulatory approval, for use in newly diagnosed and previously treated intermediate-2 or high-risk primary or secondary MF worldwide between 2019–2021.

Herein, we review the clinical trial data leading to fedratinib’s approval and collate data accruing from reports of the “real-world’ use of this drug and suggest how this drug may best be integrated into the rapidly expanding MF therapeutic armamentarium.

The authors declare no conflicts of interest regarding the material discussed in this manuscript. AD and IL performed a literature search and analyzed the articles referenced in this review. AD, OJD and MHE critically analyzed the articles included in the review and wrote the manuscript.

In 2011, a multicenter phase I trial assessed the safety, tolerability, and efficacy of fedratinib in patients with high- or intermediate-risk primary or post PV/ET MF [4]. The maximum tolerated dose was 680 mg/day. Common adverse events included nausea, vomiting, diarrhea, anemia, and thrombocytopenia, with the latter two frequently reaching grade 3–4 severity. Doses above 520 mg were associated with increased transfusion dependence over 24 weeks. Spleen volume response rates >35% (SVR35) were achieved by 39% of patients by the week 24 and by 47% at week 48. Interestingly, there was a significant decrease in the JAK2V617F allele burden after six cycles (P = 0.04) and 12 cycles of treatment (P = 0.01).

A subsequent phase II open-label study [5] randomized 31 patients to receive fedratinib 300, 400, or 500 mg once daily in 4 week cycles. Spleen and symptom responses were evaluated. The 400 mg dose showed the best spleen response, with 60% of patients responding at week 24. This study also evaluated symptom response using the standardized myeloproliferative neoplasms (MPN) total symptom score (TSS) and demonstrated that the 400 mg dose also achieved the best symptom response. The most common non-hematologic treatment-emergent adverse events (TEAEs) were gastrointestinal complaints, mainly grade 1/2. All patients experienced anemia, mostly grade 3/4, and thrombocytopenia occurred in 55%.

The JAKARTA study [6] was a phase III trial conducted across 94 centers worldwide, involving 289 previously untreated patients with high-risk or intermediate-2 risk primary or post PV/ET MF. Patients were randomized to receive either 400 mg or 500 mg of fedratinib daily, or placebo. At week 24 and 4 weeks later (the primary endpoint), SVR35 was 36% for the 400 mg group and 40% for the 500 mg group, compared to 1% for placebo. Symptom responses (defined as ≥50% reduction in MPN-TSS at week 24) were 36% and 34% for the 400 mg and 500 mg doses, respectively, versus 7% for placebo. Lower SVR35 rates were noted in patients with platelet counts <100 000/µL, higher IPSS risk status, and wild-type JAK2. The most common hematologic toxicity was anemia, reaching its nadir at 12–16 weeks, with partial recovery in the 400 mg group but not the 500 mg group. 14% of patients discontinued treatment because of adverse events within the first 24 weeks. The most frequent non-hematologic adverse events were gastrointestinal symptoms. An updated analysis of JAKARTA [7] focusing on the 400 mg dose versus placebo, reported similar SVR35 results (37% vs 36% in the initial analysis), and slightly higher symptom response rate (40% vs 36% in the initial analysis).

JAKARTA-2 [8] was an open-label, single-arm phase II study conducted alongside JAKARTA to evaluate fedratinib in patients with intermediate or high-risk primary or secondary MF who were previously treated with ruxolitinib who demonstrated intolerance or disease progression. A total of 97 patients received 400 mg daily, with dose adjustments allowed based on response or toxicity. Among the 83 assessable patients, 55 (66%) were ruxolitinib-resistant, and 27 (33%) were intolerant, primarily because of hematologic toxicity based on investigator assessment. In the original “per protocol” analysis, fedratinib achieved a SVR35 of 55% in evaluable patients and a TSS response of 26%. Efficacy varied according to the reason for ruxolitinib failure, with patients with disease progression on ruxolitinib having a lower response rate to fedratinib than those experiencing ruxolitinib toxicity (Table 1).

A subsequent re-analysis of these data using an intention-to-treat (ITT) approach applied more stringent criteria for ruxolitinib failure [9] demonstrated a SVR35 of 31% at week 24. In the stringent criteria cohort (79 patients), 30% achieved SVR35, while in a sensitivity analysis cohort (66 patients receiving six cycles of fedratinib) 36% achieved SVR35. Symptom response rates were the same in the stringent criteria and the ITT cohorts (27%), while the response rate in the sensitivity analysis cohort was 32%, suggesting potential underestimation of response. SVR was not significantly affected by the reason for prior ruxolitinib failure (Table 1) or patient age. In both analyses, the most common toxicities were anemia, thrombocytopenia, and gastrointestinal disturbances, including nausea, vomiting, and diarrhea. Anemia was observed in all patients, as a treatment-emergent adverse event (TEAE) in 50%. This rate was even higher in the sensitivity analysis cohort. Notably, 19 patients (20%) permanently discontinued fedratinib due to a TEAE.

Thrombocytopenia is associated with increased disease progression and worse survival in MF patients and is a common side effect of JAK2 inhibitors. In 2022, a subgroup analysis of the JAKARTA and JAKARTA-2 was performed to compare the efficacy and safety of fedratinib in patients with pre-treatment platelet counts of 50 000–100 000/µL to those with platelets >100,000/µL [10]. Predictably, the low platelet group had higher MF risk scores, more prior therapies, and a longer time since diagnosis. In the JAKARTA trial, the SVR rate at week 24 in the low- platelets cohort was significantly higher with fedratinib compared to placebo (36% vs0%, p < 0.01). However, there was no significant difference in SVR rate between the low-platelets and high-platelets groups (36 vs 49%, p = 0.37). Similarly, in JAKARTA2 the SVR rate at week 24 was 36% in the low‐platelets cohort and 28% in the high‐platelets cohort (p = 0.41). In JAKARTA, symptom response was higher in the high platelets subgroup (42% vs. 33%), whereas in JAKARTA-2, it was higher among low platelets patients (39% vs. 20%), these differences lacking statistical significance (Table 2). Overall, in the difficult-to-treat thrombocytopenic population with more advanced myelofibrosis, we can conclude that fedratinib is more efficacious than placebo regarding SVR, and is equally effective in both low-and high-platelets groups. Similarly, fedratinib was equally efficacious regarding symptom improvement in patients with low or high platelets as defined in the analysis. Safety was assessed in a pooled cohort of 48 low-platelets and 155 high-platelets patients. Discontinuation rates due to TEAEs were similar in both groups (15% for low-platelets and 17% for high-platelets). The low-platelets group experienced higher rates of all-grade and grade 3–4 anemia. As expected, new or worsening thrombocytopenia was more common in the low-platelets cohort (44% vs. 9%), though no serious thrombocytopenia events were reported in this group.

The FREEDOM phase 3b trial [11] studied fedratinib in patients previously treated with ruxolitinib, incorporating strategies to manage gastrointestinal adverse events (AEs) and to reduce the risk of WE. These included planned dose adjustments, proscribed use of antiemetics and antidiarrheals, dietary modification, and thiamine supplementation and monitoring. Originally planned to enroll 110 patients, the study was terminated early with only 38 patients enrolled because of difficulties encountered during the COVID-19 pandemic. Nonetheless, 71% of patients received six or more cycles of therapy, and 42% received 12 cycles or more. Of the 35 evaluable patients, 25.7% met the primary efficacy endpoint of SVR35 at week 24. Analysis using the LOCF method yielded SVR35 rates of 62% at the end of treatment. Gastrointestinal TEAEs were reported in 89.5% of patients, mostly grade ≤2, which resolved within 1–2 cycles. Transient grade 1 and 2 thiamine decreases occurred in 15.8% of patients and no thiamine decreases were reported after prophylactic thiamine supplementation was introduced.

The first analysis of the FREEDOM2 phase III trial comparing fedratinib to best available therapy (BAT) in patients previously treated with ruxolitinib was published recently [12]. Patients were randomized 2:1 to receive fedratinib or BAT, which included ruxolitinib rechallenge, red blood cell transfusion, or hydroxyurea and the primary endpoint was proportion of patients reaching SVR35 at week 24 in the ITT population. 201 patients were included: 134 received fedratinib and 67 BAT (of whom 52 received ruxolitinib). Subsequently 46 BAT patients crossed over to fedratinib per-protocol. Median follow up was 64·5 weeks (inter-quartile range 37·9–104·9). SVR35 at end of cycle 6 was seen in 48 (36%) of 134 patients receiving fedratinib versus four (6%) of 67 patients receiving BAT (p < 0·0001). At week 24, 53 (40%) of 134 patients in the fedratinib group and 8 (12%) of 67 patients in the BAT group had grade 3 or greater treatment-related adverse events, most frequently anemia (fedratinib 12 of 134 (9%); BAT 6 of 67 (9%)) and thrombocytopenia (fedratinib 16 of 134 (12%); BAT 2 of 67 (3%)). Gastrointestinal adverse events occurred more frequently in the fedratinib group compared with the BAT group, but were mostly grade 1–2 in severity and more frequent in early cycles, and were less frequent than in prior clinical trials. A total of 28 of 134 (21%) patients in the fedratinib group and 3 of 67 (4%) patients in the BAT group had reduced thiamine levels and only one of the fedratinib-treated patients had a low thiamine despite thiamine supplementation. These findings further support the use of fedratinib in the second line setting for spleen size reduction and demonstrate that gastrointestinal toxicities and thiamine deficiency are not significant barriers to treatment. Furthermore, subgroup analysis data that has been presented but not yet published indicated a higher SVR35 in patients intolerant to prior ruxolitinib (56.5%) versus those relapsed/refractory (31.5%), although this trend was not seen in the stringent JAKARTA-2 analysis [13]. Patients with lower baseline platelet counts (<100,000/µL) achieved higher SVR35 compared to those with higher platelet counts (47.1% vs. 35.3%), consistent with JAKARTA-2 results. No new safety concerns were identified: gastrointestinal AEs were mainly grade 1/2, consistent with previous studies.

The key features and efficacy outcomes of the phase II and III trials discussed above are presented in Table 3.

The ongoing German phase 2 FRACTION trial [14] is evaluating the combination of fedratinib and nivolumab in patients who have previously received any JAK inhibitor therapy. This is based on the potential of fedratinib to enhance immune responses in preclinical and clinical studies, creating a therapeutic opportunity to explore immune activation with the checkpoint inhibitor nivolumab. The primary endpoint of the FRACTION trial is the optimal response rate after 12 treatment cycles.

A multicenter Mayo Clinic experience with fedratinib in twenty-eight MF patients who were relapsed or refractory to ruxolitinib was reported [15]. The investigators assessed the efficacy, toxicity, and impact on survival. Patients had previously been treated with ruxolitinib for a median duration of 18 months, with a median dose of 15 mg twice daily. Patient features were typical for this MF population: median age was 73 years, and 54% of the subjects were women. Based on imaging, splenomegaly was observed in 86% of the patients, with a median spleen size of 23 cm. Most patients (89%) reported constitutional symptoms, and 50% were transfusion-dependent. The most frequent driver mutation was JAK2 (75%), followed by CALR (18%), MPL (4%), and 4% of patients were triple-negative. Next-generation sequencing revealed mutations in ASXL1 (40%), TET2 (20%), NRAS (20%), SRSF2 (15%), U2AF1 (10%), EZH2 (5%), and TP53 (5%) among 20 informative cases. Cytogenetic abnormalities were present in 74% of patients, classified as unfavorable (30%) or very high risk (5%). The dynamic international prognostic scoring system (DIPSS) risk was high at 43%, and 69% of evaluable patients were high/very high risk by mutation-enhanced international prognostic scoring system version 2.0 (MIPSS v2). Fedratinib was initiated at a median of 4.8 years after MF diagnosis, and administered at a median daily dose of 400 mg for a median duration of 8 months. Patients who previously required >20 mg of ruxolitinib twice daily received fedratinib for a shorter median period (4 months) than those who required lower doses of ruxolitinib (9 months). Spleen response was evaluated in 24 patients and only three patients (13%) exhibited a reduction in spleen size. Symptom improvement was more frequently reported compared to the effect on the spleen size, with one-third of patients experiencing at least a 50% reduction in their symptom scores. Either symptoms and/or spleen response was achieved in 36% of patients. Overall, the median time to response was 2.5 months and the median duration of the response was 7.8 months.

Only one transfusion-dependent patient achieved anemia response by consensus response criteria. As expected, hematological and gastrointestinal toxicity were the most prevalent adverse events. Seven (25%) patients progressed while on treatment, including one transforming into blast phase. Treatment was discontinued in 54% of patients, due to disease progression and toxicity. At a median follow-up of 11.1 months following fedratinib therapy, 32% of the cohort had died, mostly due to progressive disease (67%). No significant differences in survival among symptom/spleen responders versus non-responders were observed. In addition, duration of therapy (≥6 months versus <6 months) did not impact survival. In summary, this study reported the efficacy of fedratinib in a cohort of patients with a long history of MF (average duration of 4.8 years), who had been treated with ruxolitinib for over a year and had significant splenomegaly. In this scenario, the evidence indicates that fedratinib’s main benefit may lie in providing symptomatic relief. Its effects on spleen size reduction appear to be less significant. Furthermore, patients who do not respond to increased doses of ruxolitinib may have limited therapeutic advantages from fedratinib.

Mascarenhas et al. reported on the outcomes of spleen size, MF-specific symptoms, and hematologic responses during the first 6 months of fedratinib therapy 150 patients from community oncology practices in the United States who had previously been treated with ruxolitinib [16]. The mean age of the population was 68 years. 50% of patients had received ruxolitinib at a dose ≥20 mg twice daily and the median duration of ruxolitinib before fedratinib was 7.6 months. 43% had International Prognostic Scoring System (IPSS)/DIPSS high-risk and 37% had intermediate-2 risk disease. Fedratinib was started at 400 mg once a day in 74% of patients. The median time of follow-up post-fedratinib initiation was 5 months and the median duration of fedratinib treatment was 4.4 months. At initiation, the mean spleen length was 16.0 cm, decreasing significantly to 13.2 cm at 3 months and 7.2 cm at 6 months. The mean number of symptoms decreased significantly, and the mean platelet count increased at 3 and 6 months. Interestingly, complete resolution of palpable spleen occurred in 21% of patients. Discontinuation of fedratinib was for disease progression in 43% of cases, and 29% of patients died after discontinuation. The authors concluded that this study provides real-world evidence of the effectiveness of fedratinib when used post ruxolitinib failure in patients with intermediate- or high-risk MF, and that a longer duration of fedratinib led to greater clinical benefits in spleen size and symptomatic burden reduction. It is important to highlight that in this study, the cohort was younger and had a shorter median exposure to ruxolitinib (7.6 months) compared to the Mayo Clinic patients [14]. Moreover, the population had a relatively good clinical profile with a mean number of symptoms of only 2.5, and only 38% of patients were transfusion-dependent. Details regarding prior ruxolitinib dose were not provided and spleen response was assessed by palpation alone. These differences may explain the discrepant findings between the reports.

A study presented in 2023 at the annual meeting of the Society of Hematologic Oncology described the outcomes of a retrospective US cohort study identified using the Integra Connect PrecisionQ database [17]. The authors reported data on 150 MF patients receiving fedratinib, pacritinib, or other therapies after ruxolitinib, with at least 6 months of follow-up. Thirty of them received fedratinib. This cohort had a median age of 75 years, showed a median time on treatment of 11 months, and a median overall survival (OS) was 27.1 months. At 6 months post-index, 96.6% of patients in the fedratinib arm were alive. In their conclusions, the authors remarked that this real-world US study suggests a longer time on treatment and OS with fedratinib in post-ruxolitinib setting compared with the other options. Full publication of these data is awaited.

Another claims-based retrospective observational study aimed to evaluate the real-world impact of fedratinib on survival in the post-ruxolitinib setting, utilizing data from the US Flatiron Health national database [18]. Patients included were stratified into two groups depending on the post-ruxolitinib care received, fedratinib (n = 70) and non-fedratinib (n = 159) cohorts. The authors subsequently identified 109 patients in the non-fedratinib subgroup who discontinued ruxolitinib following the FDA approval of fedratinib in August 2019, referred to as the non-fedratinib contemporary subgroup. The subjects were followed until death, loss to follow-up (last observed clinic visit or medication administration date), or end of the study period, whichever came first. Median age at index was 71.0 and 70.0 years, in the fedratinib and non-fedratinib groups, respectively, and approximately half of patients were female (55.7 and 50.3%). The proportion of patients with post PV/ET MF was 40% in the fedratinib group and 39.0% in the non-fedratinib group, and no patients and 9.4% of patients, respectively, had evidence of blast-phase/acute myeloid leukemia at index. Among patients with an available Eastern Cooperative Oncology Group (ECOG) performance status at index, the proportion with a score of 0–1 was higher in the fedratinib group than in the non-fedratinib group (90.2 vs 74.3%). The median time from MF diagnosis to the start of ruxolitinib therapy was 7.5 months for the fedratinib cohort and only 2.1 months for the non-fedratinib contemporary subgroup. The duration of the ruxolitinib treatment for those groups was 25.3 vs 13.7 months, respectively. Patients who began fedratinib did so after a median of 55.7 months from their MF diagnosis. The median duration of fedratinib therapy in the fedratinib group was 3.7 months. Overall, 47.1% of patients receiving fedratinib started treatment with a daily dose of 400 mg, while 21.4% began with a daily dose of 200 mg. In the non-fedratinib group, the most frequently received post-ruxolitinib treatments were hydroxyurea (7.5%), a clinical study drug (8.8%), or danazol (3.8%). Median OS was not reached in the fedratinib group and was 17 months in the non-fedratinib group (p = 0.0223). In the non-fedratinib contemporary subgroup, the median OS was 12 months (p = 0.0091). In a multivariate Cox model analysis, factors that were associated with poorer OS included Charlson Comorbidity Index ≥1 (versus 0), ECOG PS score 2–4 (versus 0–1), and baseline BMI < 18.5 kg/m2 (versus 18.5–25 kg/m2), as well as time from MF diagnosis to index date of 6–9 years (versus ≤3 years). The authors concluded that fedratinib significantly improves survival compared with non-fedratinib therapy. The fedratinib group had a lower percentage of patients with an ECOG score of ≥2 at index, and a higher percentage of patients without transformation to blast-phase/acute myeloid leukemia, compared with the non-fedratinib groups. However, even after excluding these patients, the mortality remained substantially higher in the non-fedratinib group (36.1%) compared with the fedratinib group (21.4%). It should be noted that the authors were unable to assess the impact of allogeneic stem cell transplantation therapy on these populations. In this study, the median duration of fedratinib therapy was only 3.7 months, compared to a median treatment duration of 24.4 weeks (range 0.7–79.4) in the JAKARTA2 trial (8) and a median treatment duration of 28.3 weeks (range 1.6–101.3) in the FREEDOM trial [11], both in patients who have previously failed ruxolitinib therapy.

Passamonti et al described the characteristics and clinical outcomes of patients with intermediate-2 and high-risk MF who received fedratinib as second line treamtment real-world settings [19]. Patients in this cohort discontinued ruxolitinib due to refractoriness, relapse or intolerance. Those undergoing allogeneic bone marrow transplantation were excluded. A total of 196 patients were included with a median age at fedratinib initiation of 68.7 years. Most patients were male (62.8%), white (82.7%), diagnosed with primary MF (76.5%), had intermediate-2 MF in 56.1%, and with the JAK2V617F mutation in 56.1% of cases. Duration of the previous treatment with ruxolitinib was not reported. The median time from ruxolitinib discontinuation to fedratinib initiation was 1.2 months. With a median follow-up period of 13.8 months, 55.1% (n = 108) of the population remained on fedratinib treatment. Those who discontinued fedratinib had a median treatment duration of only 6.3 months. In 42% of patients, symptoms resolved completely within 6 months of starting fedratinib. Overall, 66.8% of patients had a reduction in the palpable spleen size during the same period. The median time to spleen response was 4 months. It is important to note that the study lacked details regarding the duration of MF and ruxolitinib treatment, as well as the median spleen size at the initiation of fedratinib therapy.

We recently reported on 16 patients with a physician-reported diagnosis of MF, who initiated fedratinib after receiving ruxolitinib, on a national managed access program. 62.5% of the patients were women and the median age was 77 years [20]. The median duration of prior ruxolitinib therapy was 17 months (range: 3–84) months. Before the initiation of fedratinib, the median spleen length by palpation was 15.5 cm below the costal margin (range: 4–22 cm). After 3 months the median spleen length was 13 cm below the costal margin (range: 2–21 cm). 2 patients showed minimal improvement after 6 months, while 3 patients progressed, and 2 patients showed no change in the spleen size. The spleen response did not improve after 12 months of treatment at which time the median spleen size was 19 cm below the costal margin (range: 2–30 cm). Regarding the MF-related symptoms, 43.75% (n = 7) of patients reported some improvement, 37.5% (n = 6) showed no change and 18.75% (n = 3) of the population had worsening of their symptoms. Gastrointestinal toxicity was the most frequent adverse effect of the drug and while 31% of patients died during follow up. Our observations showed that in MF patients who have failed to ruxolitinib, the therapeutic value from fedratinib may be modest, especially when exposure time to ruxolitinib was >12 months.

The key features and efficacy outcomes of the real-world data discussed above are presented in Table 4.

Fedratinib has a role in first and subsequent lines of MF treatment. In the controlled trial setting, significant spleen volume reduction and symptom improvement have been demonstrated, including in patients with platelet counts down to 50,000/µL, a difficult-to-treat population. However reports on the use of fedratinib in routine clinical practice may indicate that these promising results have not been fully recapitulated in this setting, possibly because of fedratinib being prescribed in patients with more advanced and resistant disease. Defining drug refractoriness is acknowledged as one of the most important challenges in contemporary MF management [21], and its relevance will continue to increase given the expanding therapeutic options for MF patients.

Identifying patients for whom fedratinib may be the JAK inhibitor of choice is important in an era in which the number of approved JAK inhibitors has increased rapidly and is challenging given the lack of direct between-drug comparisons [22]. Notably, momelotinib which is now approved for use in first and second line treatment has a salutary effect on hemoglobin and thus may be the preferred agent in anemic MF patients with requiring spleen size reduction or relief of systemic symptoms, while pacritinib, licensed for first and second line treatment of patients with a platelet count <50,000/µL should be considered in these severely thrombocytopenic patients. Addiitonally pacritinib has a positive effect on hemoglobin and thus may be useful in symptomatic MF patients with anemia. Given these drug characteristics, for which patients then may fedratinib be considered? We suggest that in the first line fedratinib may be useful in patients requiring JAK inhibition who have mild-moderate thrombocytopenia, namely a platelet count of 50 000–150 000/ µL. These patients would necessarily require ruxolitinib dose reduction if this drug was chosen but could receive fedrtinib in full dose allowing for more potent suppression of aberrant JAK pathway activation. When considering second line treatment a similar decision-making algorithm could be employed, with fedratinib used following ruxolitinib failure in patients in whom neither anemia nor thrombocytopenia are prohibitive. Theoretically, fedratinib may be used following momelotinib or pacritinib in the future although given the emerging preference for the use of the latter drugs in cytopenic MF, this drug sequencing may be less commonly employed by practitioners.

Currently, there are no published routine clinical practice data informing on the efficacy of fedratinib in the first line.

In summary, a more effective use of fedratinib may require its administration earlier in the disease course and in the context of second-line treatment. Timely identification of ruxolitinib failure that prompts a treatment change may enhance fedratinib’s therapeutic potential.