Adult patients
A total of 352 adults were included in the analysis: 119 and 233 patients for acute and chronic GvHD, respectively. Median age was 52 years (range, 14–75). The most frequent underlying diseases were: acute myeloid leukemia (33.88%), myelodysplastic syndrome (20.8%) and acute lymphoblastic leukemia (15.9%). Baseline characteristics of adult population are shown in table 1. Patients received a median of 2 previous lines of treatment for acute (range 0–5) and chronic GvHD (range 1–10). Remarkably, 78.2% and 98.2% of patients with acute and chronic GvHD, respectively, showed grades 3–4 or moderate-severe forms before starting ruxolitinib. Baseline characteristics of acute and chronic GvHD before the start of treatment are shown in table 2.
Efficacy
Acute GVHD. A total of 119 adult patients received ruxolitinib for refractory acute GvHD. The overall response rate (ORR) was 58.8% (70/119 patients), and 33.6% (40/119 patients) achieved complete remission. Median time to observe any response was 3 weeks from the beginning of treatment (range: 0.5–12) and median time to best response was 4 weeks (range: 1–77). By organs, responses were specifically evaluated in GI, skin and liver. Overall responses were 50% (50/100 patients), 77.2% (61/79 patients) and 63.1% (24/38 patients), respectively. CR were 25% (25/100 patients), 67% (53/79 patients) and 42.1% (16/38 patients), respectively. More detailed information is available on table 3. Remarkably, we observed no differences in terms of response rates between patients having received < 2 vs ≥ 2 (ORR 65.5% vs 53.3%, respectively; p = ns) or < 3 vs ≥ 3 (ORR 58.7% vs 61.9%, respectively; p = ns) lines of treatment. Ruxolitinib withdrawal occurred in 47.1% of patients (56/119 patients). Causes of withdrawal were: GvHD remission 30.3% (17/56 patients), GvHD progression 26.7% (15/56 patients), toxicity 16% (9/56 patients), infections 5.3% (3/56 patients) and others 21.4 (12/56 patients). Other causes included: death (n = 6), treatment intolerance (n = 3), relapse of the underlying disease (n = 1), secondary carcinoma (n = 1), unknown (n = 1). Median follow-up was 3.5 months (range: 0-71.6 months). Among responders, 12.8% (9/70 patients) lost their response with a median duration of response (DOR) of 45 days (range: 19–165) and 5.7% (4/70 patients) did not lost response but needed and additional line of treatment.
Chronic GVHD. A total of 233 adult patients were analyzed. The overall and complete response rates were 65.3% (153/233 patients) and 18.5% (43/233 patients) respectively. Median time to observe any response and best response was 2 months (range: 0.1–24) and 6 months (range: 0,25–48), respectively. By organs, overall and complete response rates in patients with scleroderma, GI or lung involvement were 69.8%, 75.9% and 37.8% and 22.4%, 24% and 1.5%, respectively. No differences were observed by lines of treatment (ORR 64.9% vs 66.4% for patients who had received < 2 or ≥ 2 prior lines, p = ns and 67.3% vs 63.6% for patients who received < 3 or ≥ 3 prior lines of treatment, p = ns; respectively. Ruxolitinib withdrawal occurred in 98/233 patients (42.1%); distribution of causes of withdrawal were: GvHD remission 30.6% (30/98 patients), GvHD progression 28.5% (28/98 patients), toxicity 18.3% (18/98 patients), infections 6.1% (6/98 patients) and others 16.3% (16/98 patients). Other causes included: relapse (n = 6), suboptimal response (n = 3), secondary malignancy (n = 2) or patient decision (n = 2), therapeutic effort limitation (n = 1), transplant (n = 1), unknown (n = 1). Median follow-up was 25.6 months (range: 0-77.3 months). Among responders, 8.5% (20/233 patients) lost their response with a median DOR of 280 days (range: 41-2258) and 5.1% (12/233 patients) responded but needed a second line of treatment to deepen their response.
Safety
Cytopenias were described in 19.3% and 19.5% of patients with acute and chronic GvHD (grade ≥ 3: 15.1% and 5.5%, respectively). Most commonly cytopenia included thrombocytopenia and anaemia. Incidence of cytomegalovirus reactivation was 49.7% for patients with acute and 30.5% for patients with chronic GvHD. Data about infections are shown in table 4. Other toxicities included hypertransaminasemia, hypertension, muscle pain and others, and were observed in 15.1% and 15.4% of patients with acute and chronic GvHD.
Taper of immunosuppression
In the acute GvHD setting, ruxolitinib withdrawal occurred in 52.9% of responding patients, and corticosteroids were stopped in 47.2% of responders. In the chronic GvHD group, these values were 40.1% and 34.8% for ruxolitinib and corticosteroids, respectively. As shown in the Fig. 1, other immunosuppressive drugs were withdrawn in a remarkably percentage of patients.
Survival, relapse and mortality
In the acute GvHD group, 61.3% of patients died (73/119 patients). Causes of mortality were: infections 50.6% (37/73 patients), GvHD progression 30.1% (22/73 patients), relapse of the underlying disease 5.4% (4/73 patients), or others 13.7% (10/73 patients). Overall survival at 6 months was 69.1% (IC95%: 55.2–79.5) vs 19.6% (CI95%:9–32) among responders vs non responders, respectively, p < 0.001 (Fig. 2). Median time of ruxolitinib treatment was 88 days (range: 5-1907) and 40 days (range: 3-314) among responders vs non responders. Globally, 7.6% of patients relapsed of their underlying malignancy (9/119 patients). Median follow-up was 3.5 months (range: 0-71.6). Event-free survival at 6 months was 48.7% (CI95%: 40.5–58.6) (Fig. 3).
In the chronic GvHD group, 24% of patients died (56/233). Causes of mortality were: infections 9.9% (23/233 patients), GvHD progression 6.4% (15/233 patients), relapse 1.7% (4/233 patients) and others 6% (14/233 patients). The 2-year OS in responders vs non-responders was 86.8% (95%CI: 79.6–91.6) vs 67.8% (95%CI: 55.5–77.4), respectively, p < 0.001 (Fig. 2). Median time of ruxolitinib treatment was 502 days (range: 41-2191) and 188 days (range: 11-2288) among responders vs non responders. Overall, 4.3% of patients relapsed of their underlying disease (10/233 patients). Event-free survival at 24 months was 79.4% (CI95%: 74.1–85.1) (Fig. 3).
Pediatric population
In the pediatric group, a total of 42 patients were included in the analysis. Median age was 6 years (range: 0–13). The most frequent underlying diseases were: acute lymphoblastic leukemia (33.3%), acute myeloid leukemia (19%), aplastic anaemia (14.3%) and primary inmunodeficiencies (14.3%). Median of prior lines of treatment were 2 (range: 1–4) and 3 (range: 1–7) for the acute and chronic GvHD groups, respectively. Similarly, to the adult population, 75.9% and 100% of patients with acute or chronic GvHD, respectively, presented grades 3–4 aGvHD or moderate-severe chronic GvHD forms before starting ruxolitinib. Baseline characteristics and GvHD score are shown in tables 1 and 2.
Efficacy
Acute GVHD group. A total of 29 pediatric patients received ruxolitinib for refractory aGVHD. The overall response rate was 82.7% (24/29 patients), and 51.7% (15/29 patients) achieved complete remission. Median time to response was 3 weeks (range: 1–8 weeks), and median time to best response was 4 weeks (range: 1–16). Overall response rates by organs were 69.5% (16/23 patients), 88.8% (24/27 patients) and 66.6% (6/9 patients) for GI, skin and liver involvement, respectively. Complete remission rates were 56.5% (13/23), 66.6% (18/27 patients) and 33.3% (3/9 patients), respectively. Results are displayed on table 3. We observed no differences in terms of response rates between patients having received < 2 vs ≥ 2 (ORR 85.7% vs 81.8%, respectively; p = ns) or < 3 vs ≥ 3 (ORR 82.3% vs 83.3%, respectively; p = ns) lines of treatment. Ruxolitinib withdrawal occurred in 72.4% patients (21/29 patients). Causes of withdrawal included: GvHD remission 61.9%, toxicity 14.2%, infections 9.5% and others 14.2% (other causes included 2 relapses and 1 PRESS syndrome). Median follow-up was 8.9 months (range: 0.4–59.5). Among responders, 3/24 patients lost their response with a median DOR of 42 days (range: 29–232). Only 1 patient needed to use an additional treatment to improve his response.
Chronic GVHD group. Thirteen patients were analyzed. Remarkably all patients responded to ruxolitinib and 23% (3/13 patients) achieved complete remission. Median time to response was 1 month (range: 0.3-6). Median time to best response was 4 months (range: 0.7–12). By organs, all patients with GI involvement reached complete remission, and all patient with scleroderma and lung involvement obtained partial response. No stable disease or progression was observed. Detailed results are shown on table 3. Only 1 patient lost his response at 232 days and another patient needed another line of treatment to improve his response. Ruxolitinib withdrawal occurred in 30.8% (4/13 patients). Causes of withdrawal among these 4 patients were: GvHD remission 50%, toxicity 25% and infections 25%. Median follow-up was 43.1 months (range: 0.9–59.6).
Safety
Cytopenias were observed in 5/29 (17.2%) and 1/13 (7.7%) of patients with acute or chronic GvHD, respectively. Grade ≥ 3 cytopenia was observed in 3/29 (3.4%) and 1/13 (7.7%) patients. Safety data concerning infections are displayed on table 4. Regarding cytomegalovirus reactivation, incidence was 34.5% and 23.1% for the acute and chronic GvHD groups. These values were 24.1% and 46.2% before ruxolitinib.
Taper of immunosuppresion
Similarly to the adult group, ruxolitinib and other immunosuppressive drugs could be stopped in a remarkably proportion of patients (Fig. 1). For the acute GvHD group, ruxolitinib and corticosteroids withdrawal occurred in the 75% and 72.7% of patients, respectively. For the chronic GvHD group, 30.7% of patients could withdraw ruxolitinib and 50%, corticosteroids. Other drugs were also reduced.
Survival, relapse and mortality
In the aGvHD group, 37.9% (11/29 patients) died. Causes of mortality were: GvHD progression 27.2% (3/11), infections 27.2% (3/11), relapse of the underlying disease 18.8% (2/11), or others 27.2% (3/11). Overall survival at 6 months in responders vs non-responders was 77.4% (CI95%: 53.6–90) vs 20% (CI95%: 0-58.2), respectively, p < 0.001 (Fig. 2). Median time of ruxolitinib treatment was 125 days (range: 51-14-29) and 34 days (range: 12–237) among responders vs non responders. Three patients relapsed within this group. Event-free survival at 6 months was 75% (CI95%: 62.7–89.7) (Fig. 3).
In the cGvHD group, 7.7% of patients died (1/13 patients). The only patient who died was due to relapse of the underlying disease, more than 2 years after initiation of ruxolitinib. Overall, 2 patients relapsed. The median time of treatment with ruxolitinib was 885 days (range: 29-1790). The two-year OS was 100% and the 2-year event free survival was 91.7% (CI95%: 77.3–100) (Fig. 3).