Phase 3 Trial Shows More Efficacy With Ruxolitinib Than Best Available Therapy in Chronic GVHD


The phase 3 REACH3 trial of ruxolitinib in patients with chronic graft-versus-host disease with an inadequate response to corticosteroids demonstrated significantly higher overall response rate, a substantially greater improvement in failure-free survival, and greater symptom improvement compared with best available therapy.

Robert Zeiser, MD

The phase 3 REACH3 trial (NCT03112603) of ruxolitinib (Jakafi) in patients with chronic graft-versus-host disease (cGVHD) with an inadequate response to corticosteroids demonstrated significantly higher overall response rate (ORR), a substantially greater improvement in failure-free survival (FFS), and greater symptom improvement compared with best available therapy (BAT), according to a presentation at the 2020 American Society of Hematology Annual Meeting.

Results demonstrated that the JAK inhibitor elicited an ORR of 49.7% in this patient population vs 25.6% with BAT (odds ratio [OR], 2.99; 95% CI, 1.86-4.80; < .0001). In the investigational arm, 11 patients (6.7%) achieved a complete response (CR), while 71 patients (43.0%) had a partial response (PR). Five patients (3.0%) in the control arm experienced a CR, while 37 patients (22.6%) had a PR.

Moreover, the median FFS also proved to be longer with ruxolitinib compared with BAT. Specifically, the median FFS was 5.7 months in the control arm and had not been reached in the investigational arm (HR, 0.370; 95% CI, 0.268-0.510; < .0001). The Modified Lee Symptom Scale (mLSS) response rate at week 24 was 24.2% with ruxolitinib arm vs 11.0% in the BAT arm (OR, 2.62; 95% CI, 1.42-4.82; = .0011). Thus, those who received the JAK inhibitor experienced a greater improvement in symptoms.

“This is the first successful randomized phase 3 trial in adolescent and adult patients with cGVHD with an inadequate response to corticosteroids,” Robert Zeiser, MD, professor of hematology and oncology at the University Medical Center Freiburg, said during a virtual presentation on the data delivered as part of the first press program. “Ruxolitinib is the first agent to demonstrate superior efficacy to BAT in a phase 3 trial [in this patient population].”

Approximately 30% to 70% of patients who undergo allogeneic stem cell transplant will experience cGVHD, and this is the leading cause of nonrelapse mortality in these patients, according to Zeiser. Frontline treatment for this population is typically comprised of systemic corticosteroids, but only half of patients will respond to this approach. The other 50% of patients will become either refractory to or dependent on corticosteroids. No second-line standard of care has been established for these patients.

The randomized phase 3 REACH3 trial enrolled patients who were aged 12 years or older and had steroid-refractory or -dependent cGVHD. This was defined as: not responding to prednisone given at a daily dose of 1 or more mg/kg for 1 week or longer or experiencing progressive disease on the treatment; having persistent disease that does not improve with prednisone delivered at a daily dose of greater than 0.5 mg/kg or 1 mg/kg delivered every other day for 4 weeks or longer; or a prednisone dose increase to greater than 0.25 mg/kg per day following 2 unsuccessful efforts to taper the dose. Patients also needed to have evidence of myeloid and platelet engraftment.

Participants were randomized 1:1 to receive either ruxolitinib at 10 mg twice daily (n = 165) or BAT (n = 164).

“Patients who were on corticosteroids were allowed to continue that alone or in combination with a calcineurin inhibitor,” noted Zeiser.

The primary efficacy period was under week 24. The primary end point, which was ORR using National Institutes of Health consensus criteria for response, was evaluated at this time. After this time point, patients from the BAT arm were permitted to cross over to ruxolitinib. The trial also had an extension period, which was followed by a safety follow-up period. Key secondary end points included FFS and mLSS response at week 24.

Patient characteristics at baseline were reported to be well balanced between the 2 arms with regard to age (median 49.0 years vs 50.0 years, respectively), gender (66.1% male vs 56.1%), and cGVHD severity (moderate: 41.2% vs 44.5%; severe: 58.8% vs 54.9%).

Additionally, the median time from cGVHD onset to randomization in the ruxolitinib arm was similar to that of the BAT arm, at 24.9 weeks versus 21.4 weeks, respectively. Stem cell source, donor type, and cytomegalovirus (CMV) status were also found to be well matched between the arms. For most patients, peripheral blood served as the stem cell source, slightly more patients had a donor type that was related, and a little less than half of patients in each arm had a donor/recipient CMV status of positive/positive.

Notably, more patients on ruxolitinib remained on treatment at the time of the primary analysis compared with those on BAT. Specifically, 50.3% (n = 83) of patients were still receiving ruxolitinib, while only 25.6% (n = 42) of patients were still receiving BAT. Less patients on ruxolitinib discontinued treatment compared with those on the BAT arm, at 49.7% (n = 82) vs 74.4% (n = 122), respectively.

Toxicities served as the reason for discontinuation in 17.0% (n = 28) of patients on ruxolitinib arm vs 4.9% of those on BAT. However, a lack of efficacy was a major cause of discontinuation in the BAT arm versus the ruxolitinib arm, at 42.7% (n = 70) vs 14.5% (n = 24), respectively. Other notable reasons included physician decision (ruxolitinib: 2.4% vs BAT: 8.5%) and patient or guardian decision (2.4% vs 6.7%, respectively). Additionally, 37.2% of patients in the BAT arm crossed over to the ruxolitinib arm.

Additional results showed that the best ORR was also higher with the JAK inhibitor compared with BAT, at 76.4% vs 60.4%, respectively. In the investigational arm, the CR rate was 12.1% and the PR rate was 64.2%; in the control arm, these rates were 6.7% and 53.7%, respectively. The median duration of best overall response was 6.24 months with BAT and had not yet been reached with ruxolitinib.

Any-grade adverse effects (AEs) occurred in 97.6% (n = 161) of patients who received ruxolitinib versus 91.8% (n = 145) of those given BAT. Grade 3 or greater toxicities were experienced by 57.0% (n = 94) of those on the ruxolitinib arm vs 57.6% (n = 91) of those on BAT. Rates of serious AEs also proved to be similar between the arms, at 33.3% (n = 55) and 36.7% (n = 58), respectively.

“Safety is an important concern in patients who are very susceptible, particularly to infections, and who also have cGVHD and are treated with corticosteroids,” Zeiser emphasized. “We observed that any-grade AEs were not significantly different in the 2 arms; severe grade 3 or greater toxicities were also similar.”

The rate of those who experienced toxicities that resulted in dose modifications proved to be higher with ruxolitinib vs BAT, at 37.6% (n = 62) vs 16.5% (n = 26), respectively, as were the rate of AEs that led to discontinuation, at 16.4% (n = 27) vs 7.0% (n = 11), respectively. The rate of deaths up to data cutoff were similar in the 2 arms (18.8% vs 16.5%).

The most frequently reported AEs with ruxolitinib and BAT included anemia (any-grade, 29.1% vs 12.7%, respectively; grade 3 or higher, 12.7% vs 7.6%) and thrombocytopenia (any-grade, 21.2% vs 14.6%; grade 3 or higher, 15.2% vs 10.1%).

“No significant difference in viral or bacterial infections was observed between the arms,” said Zeiser. “However, we did see a trend toward a higher frequency of fungal infections [with ruxolitinib].”

CMV infection or reactivation was 5.5% in patients who received ruxolitinib vs 8.2% in those who received BAT. Overall, the toxicity profile of ruxolitinib proved to be consistent with prior data and no new safety signals were observed.


Zeiser R. Ruxolitinib vs best available therapy in patients with steroid-refractory/steroid-dependent chronic graft-vs-host disease: primary findings from the phase 3, randomized REACH3 study. Presented at: 2020 ASH Annual Meeting; December 5-8, 2020; Virtual. Abstract 77.

Related Videos
Related Content