Long-term safety and effectiveness of ruxolitinib in patients with myelofibrosis in Japan: an observational study

ABSTRACT

Aim

To evaluate the safety and effectiveness of ruxolitinib in patients with myelofibrosis (MF) in Japan.

Methods

A multicenter, observational study of patients who received ruxolitinib for MF from July 2014.

Results

Of 892 patients (mean age: 70 years, 45.9% primary MF, ruxolitinib treatment median duration, 541.0 days), 67.7% had adverse drug reactions (ADRs) and 31.5% had serious ADRs. The most frequent ADRs were anemia and decreased platelet count. Incidences of ADRs by time of onset were 57.7%, 20.3%, 14.4%, 11.1%, 11.3%, 9.0%, and 1.8% from the treatment initiation to Day 182, and every 6 months thereafter until Day 1,093 or later, respectively. ADRs of special interest included myelosuppression (46.8%), infections (17.6%), hepatic impairment (13.5%), hemorrhagic events (10.2%), cardiac failure (2.5%), interstitial lung disease (1.5%), malignancy (1.4%) and tuberculosis (0.5%). Incidences of common ADRs were similar between patients with hepatic or renal impairment and patients without hepatic or renal impairment. At 6 months, spleen responses and symptom improvement were observed in 26.2% and 52.0% of patients, respectively. Median overall survival was not reached.

Conclusion

In a real-world setting in Japan, ruxolitinib demonstrated a reasonable degree of effectiveness with no new safety concerns. Results were similar to those from clinical trials.

Plain Language Summary

Myelofibrosis (MF) is a rare type of blood cancer that interferes with the process of blood cell production by the bone marrow. In patients with MF, the bone marrow becomes overactive, leading to scarring, and subsequently, there is a lack of healthy blood cells being produced. The main symptoms of MF include anemia, fatigue, weakness and pain or discomfort in the abdomen.Ruxolitinib became available in Japan as a MF treatment in 2014. In this study, we explored the safety and effectiveness of ruxolitinib in real-world settings in Japan. We assessed 892 patients who had MF and received ruxolitinib. Most patients (72.9%) were aged 65 years or older. Side effects (adverse drug reactions, ADRs) were reported in 67.7% of patients, and 31.5% had serious ADRs. The most common ADRs were anemia (27.8%) and low blood platelet counts (21.4%). Most of the ADRs were seen in the first 6 months of starting ruxolitinib and then became less frequent over time. Patients with liver or kidney problems had a similar number of ADRs as that in patients without these problems. After 6 months of ruxolitinib treatment, 26.2% of patients had a spleen response and 52.0% of patients had less symptoms. We could not calculate the median overall survival because the observation period ended before it was achieved. In a Japanese real-world clinical setting, our results did not show any new safety concerns and were similar to those reported in clinical trials.

1. Introduction

Myelofibrosis (MF) is a chronic, progressive, disorder of the bone marrow that can occur on its own (primary) or develop from myelofibrotic progression of essential thrombocythemia or polycythemia vera (secondary) [1]. MF is a myeloproliferative neoplasm (MPN) with a clinical phenotype characterized by significant enlargement of the spleen, worsening anemia, and constitutional symptoms, which significantly impair patients’ quality of life [2]. Although the clinical course of patients with MF can vary, survival is generally limited as death can occur due to complications such as leukemic transformation, infection, bleeding, and portal hypertension [1,2].

Patients with MF are typically aged over 65 years and present with unexplained abnormalities in blood counts, often experiencing a substantial burden of symptoms [3]. Epidemiological data on MPNs are limited; however, incidence rates of primary MF range from 0.3 to 0.8 per 100,000 person-years in Western countries [3]. In Japan, the incidence of both primary and secondary MF is estimated at 0.2 per 100,000 person-years, with a prevalence of 1.0 per 100,000 person-years [4]. In Japan, the median age at diagnosis is 66 years, and the median survival duration is 47 months with a 3-year survival rate of 59% [5].

Allogeneic hematopoietic stem-cell transplant (HSCT) is the only curative treatment currently available for MF. However, it is associated with substantial morbidity and mortality, and thus it is considered an option only for patients with MF who lack significant comorbidities, possess high-risk molecular features, and have an available donor [6]. For patients with MF who cannot undergo HSCT, the main treatment goals are symptom control and prolonged survival [7]. The development of Janus kinase (JAK) inhibitors has transformed treatment prospects for patients with MF. Ruxolitinib, a JAK 1/2 inhibitor, was approved by the United States Food and Drug Administration for the treatment of MF in 2011 [8], and has since been approved in more than 100 countries. Clinical trials have demonstrated that ruxolitinib treatment prolongs survival, reduces symptom burden, reduces spleen volume, and improves quality of life in patients with MF [9–11]. Several major guidelines, including those from the Japanese Society of Hematology, recommend ruxolitinib as a crucial treatment for improving patients’ quality of life and managing symptoms in MF [12,13].

Ruxolitinib was approved for the treatment of MF in Japan in 2014 [14], primarily based on data from the phase 3 COMFORT clinical trials [9,10] and an Asian multinational phase 2, open-label, single-arm trial (Study A2202) [15]. Since data from Japanese patients in these clinical trials were limited, an all-case surveillance study was set as a condition for this approval until post-marketing data could be accumulated for a specified number of patients. This report presents the results of this post-marketing surveillance (PMS), which assessed the long-term safety and effectiveness of ruxolitinib treatment for patients with MF in real-world settings in Japan.

2. Patients and methods

2.1. Study design

This multicenter, non-interventional study was conducted in a real-world setting among patients with MF in Japan. The study was mandated by the Pharmaceuticals and Medical Devices Agency and was conducted in accordance with the Good Post-Marketing Study Practice for Drugs (GPSP) as ordained by the Ministry of Health, Labour and Welfare [16]. After ruxolitinib was approved for MF treatment in Japan on 4 July 2014, an all-case surveillance study was initiated.

This study included all patients who received ruxolitinib for the treatment of MF on or after 4 July 2014. The observational period lasted for 1 year after the first dose of ruxolitinib and extended up to 3 years if the patient continued ruxolitinib for more than one year. Patients who had previously received ruxolitinib or received ruxolitinib for the treatment of polycythemia vera (PV) were excluded. As a non-interventional study, there were no specifications concerning the treatment plan, diagnostic or treatment procedures, or visit schedule. The frequency of visits and assessments followed the typical routine medical care. For each patient, investigators completed case report forms (CRFs) documenting patient demographics, disease characteristics, ruxolitinib administration, concomitant medication, non-drug therapies for MF, spleen length, physician’s global assessment of improvement, laboratory tests, and adverse events (AEs).

2.2. Safety assessments

AEs were identified through physical examinations, patient reports, laboratory values, and other tests conducted during the study period. All AEs were coded in accordance with the Japanese version of the Medical Dictionary for Regulatory Activities (MedDRA/J) version 25.1 and classified and graded according to the Common Terminology Criteria for Adverse Events (CTCAE) version 4.0. AEs were documented on the CRF, in addition to the onset of the AE, AE term, CTCAE grade, causal relationship to ruxolitinib, duration of AE, outcome, and whether the AE was considered a serious AE (SAE). Any events that were not initially deemed serious but were later considered serious based on further details – as well as any events resulting in death – were classified as SAEs. AEs that could not be ruled out as being unrelated to ruxolitinib based on the descriptions in the CRF were regarded as adverse drug reactions (ADRs). ADRs of special interest in this study included myelosuppression, infections, tuberculosis, hepatic impairment, hemorrhagic events, cardiac failure, interstitial lung disease, malignancy (secondary carcinoma), and ADRs in patients with hepatic impairment or renal impairment.

2.3. Effectiveness assessments

For spleen size, length of the greatest spleen protrusion from the costal margin by palpation was measured before the start of ruxolitinib treatment and after 6 months and 1 year of treatment, or at treatment discontinuation. For patients receiving ruxolitinib for more than 1 year, spleen size was measured at 2 and 3 years, or upon treatment discontinuation. When the spleen was not palpable, the length was recorded as 0 cm. Physicians evaluated patients for symptom relief through Physician Global Assessments (PGA). These assessments were scheduled at 6 months and 1 year, or upon treatment discontinuation. For patients who continued ruxolitinib beyond 1 year, subsequent PGAs were scheduled at Years 2 and 3, or upon treatment discontinuation. Overall survival was assessed at 6-, 12-, 24-, and 36-months from the start of ruxolitinib treatment.

2.4. Statistical analyses

The target number for the enrollment of patients with the 1 year follow-up period was 180. Descriptive statistics were generated for all continuous, categorical, and binary study variables; the data were presented as means, standard deviations (SD), medians and minimum and maximum values for continuous variables and as frequencies and percentages for categorical variables.

The safety analysis set included all patients whose data were considered valid for analysis and excluded those patients for whom the CRF lacked a physician signature, where ruxolitinib administration was outside the contract period, patients were not enrolled or had unconfirmed enrollment, duplicate patients, off label use, patients who were not treated with ruxolitinib, and patients with prior ruxolitinib treatment. The safety analysis period spanned from the start date of ruxolitinib treatment until the end date of the observation period (or the date of ruxolitinib discontinuation) plus 28 days. Events that occurred outside the safety analysis period were excluded from the analysis.

AE analyses included the proportion of patients with AEs and SAEs by MedDRA system organ class (SOC) and preferred term (PT), the proportion of patients with AEs leading to ruxolitinib discontinuation, and AEs leading to death. The occurrence of multiple AEs with the same MedDRA PT or within the same SOC for a single patient was regarded as one patient or event. ADRs were analyzed as per AEs. Additionally, the proportion of patients with ADRs in specific categories of patient characteristics was analyzed. Odds ratios of the ADRs were calculated for patient characteristics based on the following categories: sex (male or female), age ( <15 years, ≥15 to < 65 years, and ≥65 years), duration of disease (< 1 year, ≥1 year to < 3 years, ≥3 years to < 5 years, ≥5 years to < 10 years, and ≥10 years), presence of JAK mutations, platelet count at the start of treatment ( < 100,000/mm³, ≥100,000/mm³, < 100,000/mm³, ≥100,000/mm³ to < 200,000/mm³, and ≥ 200,000/mm³), Eastern Cooperative Oncology Group performance status (0, 1, 2, 3, and 4), presence of comorbidities (hepatic impairment, renal impairment, tuberculosis, and infections) and concomitant drug status (drugs for the treatment of primary disease, prophylactic drugs for opportunistic infections, hematopoietic growth factors, and CYP3A4 inhibitors). Adjusted odds ratios for the ADRs were then calculated using a multivariate logistic regression model with the patient factors whose 95% confidence intervals (CI) of the unadjusted odds ratio did not include 1.

The effectiveness analysis set included patients who met the inclusion criteria for the safety analysis set and excluded those for whom effectiveness was indeterminable. The proportion of patients, along with 95% CI, was calculated for each spleen shrinkage category (complete disappearance, ≥50%, ≥30 to < 50%, < 30%, unchanged, and increased), PGA improvement category (remarkable improvement, improvement, no change, exacerbation, could not be determined, or unknown/not specified), spleen responders (patients with complete disappearance or a ≥ 50% reduction from baseline in spleen size) and PGA responders (patients assessed as having remarkable improvement or improvement). Odds ratios of the PGA response rates were calculated by patient characteristics. Adjusted odds ratios were then calculated using a multivariate logistic regression model with patient factors whose 95% CI of the unadjusted odds ratio did not include 1. Kaplan – Meier curves for overall survival were plotted and 6-, 12-, 24-, and 36-month probabilities of survival with 95% CI were calculated using Greenwood’s formula. Additionally, Kaplan-Meier curves for overall survival were analyzed by patient characteristics.

All statistical analyses were conducted using SAS version 9.4. There was no imputation for missing data.

3. Results

3.1. Patient disposition

A total of 2914 patients with MF were enrolled. After excluding patients who were enrollment only or ineligible for this analysis, 892 and 891 were included in the safety and effectiveness analysis sets, respectively (Figure 1).

Figure 1.

Patient disposition.

3.2. Patient characteristics

For the 892 patients included in the safety analysis population, the median age at baseline was 70.0 years, with 72.9% (650/892) of patients being 65 years or older (Table 1). Two (0.2%) patients were younger than 18 years (one patient was 3 years old, and the other was 16 years old). The mean duration of disease was 4.5 years and 39.8% (355/892) of patients had a disease duration of less than 1 year. The reasons for ruxolitinib treatment were primary MF (409/892, 45.9%), MF progressing from essential thrombocythemia (251/892, 28.1%), MF progressing from PV (197/892, 22.1%), and MF from other causes (33/892, 3.7%).

Table 1.

Patient demographic and clinical characteristics.

	N = 892
Male, n (%)	491 (55.0)
Age, years
Mediana (range)	70.0 (3–92)
< 15 years	1 (0.1)
≥15 and <65 years	236 (26.5)
≥65 years	650 (72.9)
Duration of disease, years
Mean (SD)	4.5 (6.2)
<1 year	355 (39.8)
≥1 year and <3 years	130 (14.6)
≥3 years and <5 years	74 (8.3)
≥5 years and <10 years	101 (11.3)
≥10 years	139 (15.6)
Unknown	93 (10.4)
Reason for ruxolitinib treatment, n (%)
Primary MF	409 (45.9)
MF progressing from PV	197 (22.1)
MF progressing from ET	251 (28.1)
Other MF	33 (3.7)
Unknown	2 (0.2)
JAK mutations, n (%)b	462 (51.8)
Prognostic factors
Age ≥65 years	650 (72.9)
Persisting pyrexia, night sweats, and weight decreased	213 (23.9)
Hb <10 g/dL	593 (66.5)
WBC > 25,000/mm3	236 (26.5)
Peripheral bone marrow blasts ≥ 1%	441 (49.4)
Platelet count at the start of treatment
≥100,000/mm3	624 (70.0)
<100,000/mm3	225 (25.2)
Unknown	43 (4.8)
ECOG Performance Status
0	319 (35.8)
1	366 (41.0)
2	129 (14.5)
3	66 (7.4)
4	12 (1.4)
Medical history, n (%)	315 (35.3)
Hepatic impairment	11 (1.2)
Renal impairment	4 (0.5)
TB	7 (0.8)
Infections	92 (10.3)
Complications, n (%)	590 (66.1)
Hepatic impairment	71 (8.0)
Renal impairment	71 (8.0)
TB	2 (0.2)
Infections	59 (6.6)
Received prior medication, n (%)	546 (61.2)
Primary disease treatment	514 (57.6)
Suspected to be the cause of AEs	37 (4.2)
Opportunistic infection prophylaxis	37 (4.2)
Hematopoietic growth factors	8 (0.9)
CYP3A4 inhibitors	22 (2.5)

Safety analysis set. ^aN = 887; ^bNot measured in 234 patients. Medical history is defined as conditions that are not ongoing at the time of treatment initiation. Complications are defined as conditions that are ongoing at the time of treatment initiation. Patients being treated with ruxolitinib for other MF refers to the patients who are having secondary MF or MF progressing from disease other than polycythemia vera or essential thrombocythemia, such as myelodysplastic syndromes or aplastic anemia.

Abbreviations: AE, adverse event; CYP, cytochrome P450; ECOG, Eastern Cooperative Oncology Group; ET, essential thrombocythemia; Hb, hemoglobin; JAK, Janus kinase; MF, myelofibrosis; PV, polycythemia vera; SD, standard deviation; TB, tuberculosis; WBC, white blood cell.

3.3. Ruxolitinib treatment

The mean (SD) daily dose of ruxolitinib was 17.8 (10.6) mg. The distribution of patients based on the initial daily dose of ruxolitinib was as follows: < 10 mg (126/892, 14.1%), ≥10 mg and < 20 mg (413/892, 46.3%), ≥20 mg and < 30 mg (175/892, 19.6%), ≥30 mg and < 40 mg (92/892, 10.3%), ≥40 mg and ≤50 mg (85/892, 9.5%), and > 50 mg (1/892, 0.1%). Apart from one patient who received a dose above 50 mg, ruxolitinib was administered within the approved dosage and administration ranges. Mean (SD) duration of ruxolitinib treatment was 587.0 (437.3) days and most (395/892, 44.3%) patients received ruxolitinib for a duration of 729–1092 days (Table S1).

Of the 892 patients included in the safety analysis set, 64.0% (n = 571) discontinued the study (Table S2). The most frequent reasons for treatment discontinuation were AEs (237/892, 26.6%), exacerbation of primary disease (139/892, 15.6%) and hospital transfer (93/892, 10.4%). The discontinuation rate by the timing of treatment was lowest from the start of treatment to Day 14 (31/892, 3.5%) and ranged between 11.4% and 25.6% from Day 15 to the end of the study.

3.4. Safety outcomes

The incidences of AEs and SAEs were 84.8% (756/892) and 53.3% (475/892), respectively. The most common AEs and SAEs included anemia, decreased platelet count, MF, primary MF, and pneumonia (Table 2).

Table 2.

Incidence of AEs and ADRs.

n (%)	AEs	SAEs	ADR	Serious ADR
Patients with events	756 (84.8)	475 (53.3)	604 (67.7)	281 (31.5)
Most frequent eventsa
Anemia	297 (33.3)	104 (11.7)	248 (27.8)	80 (9.0)
Platelet count decreased	231 (25.9)	67 (7.5)	191 (21.4)	48 (5.4)
MF	125 (14.0)	70 (7.9)	22 (2.5)	9 (1.0)
Primary MF	104 (11.7)	66 (7.4)	12 (1.4)	7 (0.8)
Pneumonia	73 (8.2)	62 (7.0)	37 (4.2)	33 (3.7)

Safety analysis set (N = 892). ^a ≥5% in any category. The occurrence of multiple events with the same preferred term in a patient was counted as one event. MedDRA/J version 25.1.

Abbreviations: ADR, adverse drug reaction; AE, adverse event; MF, myelofibrosis; SAE, serious adverse event.

ADRs occurred in 67.7% (604/892) of patients and serious ADRs occurred in 31.5% (281/892) of patients (Table 2). The most common ADRs and serious ADRs were anemia (27.8% [248/892] and 9.0% [80/892], respectively) and decreased platelet count (21.4% [191/892] and 5.4% [48/892], respectively). Thrombocytopenia and decreased platelet counts were collected as separate categories per the treating investigator and coded as such; therefore, patients may have experienced an AE in each of these two categories. ADRs of thrombocytopenia were reported in 16/892 (1.8%) patients with 4/892 (0.5%) patients categorized as serious ADRs. The incidence of ADRs by the time of onset was 57.7% (515/892), 20.3% (132/652), 14.4% (78/542), 11.1% (51/461), 11.3% (46/408), 9.0% (33/365), and 1.8% (6/330) from the start of treatment to Day 182, and every 6 months thereafter until Day 1093 or later, respectively (Table S3). Most ADRs had occurred by Day 182 and prolonged use of ruxolitinib did not noticeably increase the incidence of ADRs. Common ADRs (≥5%) that occurred on Day 181 or later were anemia (42/654; 6.4%) and platelet count decreased (33/654; 5.1%), similar to those occurring by Day 182.

Owing to AEs, 32.1% (286/892) of patients discontinued ruxolitinib, while 16.0% (143/892) discontinued due to ADRs. Major AEs leading to treatment discontinuation (≥2%) were MF (43/892, 4.8%), primary MF (42/892, 4.7%), anemia (30/892, 3.4%), and decreased platelet count (29/892, 3.3%) and major ADR leading to treatment discontinuation were anemia (26/892, 2.9%) and platelet count decreased (23/892, 2.6%).

AEs with a fatal outcome were reported in 25.9% (231/892) of patients; the most frequently reported AEs leading to death were primary MF (52/892, 5.8%), MF (50/892, 5.6%), pneumonia (20/892, 2.2%), sepsis (10/892, 1.1%) and leukemia (9/892, 1.0%). ADRs leading to death occurred in 7.0% (62/892) of patients; the most common ADRs were pneumonia (8/892, 0.9%) and MF (8/892, 0.9%).

The ADRs of special interest (and the proportions of patients), were myelosuppression (417/892, 46.8%), infections (157/892, 17.6%), hepatic impairment (120/892, 13.5%), hemorrhagic events (91/892, 10.2%), cardiac failure (22/892, 2.5%) interstitial lung disease (13/892, 1.5%), malignancy (secondary carcinoma; 12/892, 1.4%) and tuberculosis (4/892, 0.5%; Table S4). The most frequently reported secondary malignancies were lung neoplasm malignant (2/892, 0.2%) and squamous cell carcinoma of skin (2/892, 0.2%); all other secondary malignancies (including basal cell carcinoma, carcinoid tumor of the gastrointestinal tract, gastric cancer lung adenocarcinoma, second primary malignancy, uterine cancer, neuroendocrine tumor, lung neoplasm malignant, prostate cancer and gingival cancer were reported in 0.1% [1/892] of patients each). The ADRs of special interest among the 71 patients with hepatic impairment, 76.1% (n = 54) had an ADR, and of the 71 patients with renal impairment, 83.1% (n = 59) had an ADR (Table S5). The type and incidence of ADRs were similar in patients with and without hepatic impairment, with the most frequently reported ADRs being anemia (31.0% [22/71] and 27.5% [226/821], respectively) and platelet count decreased (18.3% [13/71] and 21.7% [178/821], respectively; Table S5). Similarly, the most frequently observed ADRs in patients with and without renal impairment were anemia (32.4% [23/71] and 27.4% [225/821], respectively) and platelet count decreased (19.7% [14/71] and 21.6% [177/821], respectively; Table S5).

In an analysis of the incidence of ADRs by patient factors, the 95% CI of the unadjusted odds ratio for age, platelet count at the start of treatment, and the presence or absence of renal impairment did not include 1 (Table S6). Likewise, the 95% CI of the adjusted odds ratio for these factors from a multivariate logistic regression analysis did not include 1. The incidence of ADRs by age was 60.6% (143/236) and 70.3% (457/650) in patients aged ≥ 15 to < 65 years and ≥65 years, respectively; the unadjusted odds ratio (95% CI) was 1.54 (1.13–2.10) and the adjusted odds ratio (95% CI) was 1.53 (1.11–2.12). Patients with a platelet count of < 100,000/mm³ at the start of treatment had a lower incidence of ADRs (61.8%, 139/225) than those with counts ≥ 100,000/mm³ (70.7%, 441/624); the unadjusted odds ratio (95% CI) was 0.67 (0.49–0.92) and the adjusted odds ratio was 0.70 (0.51–0.97). A difference was observed in the incidence of anemia, which was lower in patients with a platelet count of < 100,000/mm³ (13.3%, 30/225) than in those with a count of ≥ 100,000/mm³ (33.7%, 210/624). As for renal impairment, the incidence of ADRs was higher in patients with renal impairment (59/71, 83.1%) than in those without (545/821, 66.4%); the unadjusted odds ratio [95% CI] was 2.49 (1.32–4.71) and the adjusted odds ratio (95% CI) was 2.30 (1.18–4.49).

3.5. Effectiveness

Among the 891 patients in the effectiveness analysis set, spleen size measurements were available for 319 patients at baseline and after the start of treatment. The mean (SD) spleen size at the start of treatment was 12.0 (7.4) cm, and 8.2 (7.0) cm at 6 months, 7.7 (6.5) cm at 12 months, 6.9 (5.9) cm at 24 months, 7.2 (5.4) cm at 36 months, and 9.5 (8.6) cm at the final evaluation. Among the 286 patients with a spleen measurement of > 0 cm at baseline, the mean (SD) of percent change from baseline was − 32.0% (36.2), −34.9% (41.5), −37.4% (46.7), −31.6% (44.7), and − 18.8% (63.4) at 6, 12, 18, 24, and 36 months, and at the final evaluation respectively. Among the 319 patients (including those for whom the spleen was not palpable and was calculated to be 0 cm), the proportion of patients with a spleen response (defined as complete disappearance or a ≥ 50% reduction in spleen size from baseline) was 26.2% (54/206), 33.1% (43/130), 38.1% (32/84), 30.6% (15/49), and 22.6% (72/319) at 6, 12, 24, and 36 months, and at the final evaluation, respectively (Table 3). Of note, the 33 patients with a baseline spleen size of 0 cm were only included in the evaluation categories of unchanged or increased (Table 3).

Table 3.

Proportion of patients with spleen shrinkage.

	Complete disappearance n (%)	≥50% n (%)	30% to 50% n (%)	< 30% n (%)	Unchanged n (%)	Increased n (%)	Responders n/N	Response rate (95% CI)
After 6 months	17 (8.3)	37 (18.0)	32 (15.5)	56 (27.2)	50 (24.3)	14 (6.8)	54/206	26.2 (20.4–32.8)
After 12 months	14 (10.8)	29 (22.3)	23 (17.7)	27 (20.8)	22 (16.9)	15 (11.5)	43/130	33.1 (25.1–41.9)
After 24 months	13 (15.5)	19 (22.6)	11 (13.1)	15 (17.9)	14 (16.7)	12 (14.3)	32/84	38.1 (27.7–49.3)
After 36 months	5 (10.2)	10 (20.4)	8 (16.3)	8 (16.3)	5 (10.2)	13 (26.5)	15/49	30.6 (18.3–45.4)
At the final evaluation	28 (8.8)	44 (13.8)	48 (15.1)	83 (26.0)	58 (18.2)	58 (18.2)	72/319	22.6 (18.1–27.6)

Efficacy analysis set (N = 891). Patients with a complete disappearance or ≥ 50% were classed as responders.

Abbreviations: CI, confidence interval; n, number of patients in the specified category; N, total number of patients assessed.

Among the 891 patients in the effectiveness analysis set, the response rate of PGA improvement in symptoms (defined as remarkable improvement or improvement) was 52.0% (463/891), 48.8% (297/609), 39.0% (201/516), 35.4% (139/393) and 30.3% (270/891) at 6, 12, 24, and 36 months, and at the final evaluation, respectively (Table 4).

Table 4.

PGA of improvement in symptoms.

	Remarkable improvement n (%)	Improvement n (%)	No change n (%)	Exacerbation n (%)	Could not be determined n (%)	Unknown/ not specified n (%)	Responders n/N	Response rate (95% CI)
After 6 months	70 (7.9)	393 (44.1)	229 (25.7)	33 (3.7)	166 (18.6)	0	463/891	52.0 (48.6–55.3)
After 12 months	40 (6.6)	257 (42.2)	213 (35.0)	30 (4.9)	67 (11.0)	2 (0.3)	297/609	48.8 (44.7–52.8)
After 24 months	26 (5.0)	175 (33.9)	223 (43.2)	43 (8.3)	48 (9.3)	1 (0.2)	201/516	39.0 (34.7–43.3)
After 36 months	24 (6.1)	115 (29.3)	178 (45.3)	41 (10.4)	34 (8.7)	1 (0.3)	139/393	35.4 (30.6–40.3)
At the final evaluation	40 (4.5)	230 (25.8)	352 (39.5)	101 (11.3)	164 (18.4)	4 (0.5)	270/891	30.3 (27.3–33.4)

Effectiveness analysis set. Patients with remarkable improvement or improvement were categorized as responders.

Abbreviations: CI, confidence interval; n, number of patients in the specified category; N, total number of patients assessed; PGA, physician global assessment.

Odds ratios and 95% CIs were calculated for patient characteristics using the same categories as those used for the ADR analysis. The 95% CI of the unadjusted odds ratio did not include 1 for four factors: the duration of disease (≥1 to < 3 years and ≥ 5 to < 10 years), platelet count at treatment initiation, the presence or absence of concomitant medication for the treatment of primary disease, and prophylaxis of opportunistic infections (Table S7). The adjusted odds ratio and 95% CI for these characteristics were calculated using a multivariate logistic regression model. The 95% CI of the duration of disease and the presence or absence of concomitant medication for the treatment of primary disease did not include 1. No consistent trend was observed between the duration of disease and response rate, which was lower in patients who had MF for (≥1 to < 3 years (20.2% [26/129]; unadjusted odds ratio [95% CI]: 0.59 [0.36–0.95]) and higher in those who had MF for ≥ 5 to < 10 years (43.6% [44/101]; unadjusted odds ratio [95% CI]: 1.79 [1.14–2.82]), as compared with those who had MF for less than 1 year (30.1% [107/355]). Adjusted odds ratios (95% CI) were 0.58 (0.35–0.96) in the ≥ 1 to < 3 years group and 1.87 (1.17–2.99) in the ≥ 5 to < 10 years group. A decrease in the response rate for patients receiving concomitant medications for the primary disease was observed compared with that in those who were not receiving medications (27.3% [138/506] vs 34.3% [132/385]; unadjusted odds ratio [95% CI]: 0.72 [0.54–0.96]). The adjusted odds ratio (95% CI) was 0.65 (0.47–0.90). No concerns necessitating special precautions were identified.

Median overall survival was not reached at the end of the observation period. The estimated rates of overall survival were 92.2%, 87.1%, 78.4%, and 71.8% at 182 days, 364 days, 728 days, and 1092 days after the start of ruxolitinib treatment, respectively (Figure 2). Kaplan – Meier curves for overall survival differed among the subgroups of International Prognostic Scoring System (IPSS) risk category, initial platelet count, and initial hemoglobin level (Figure 2). Survival rates after initiating ruxolitinib at various time intervals revealed that highrisk patients had consistently lower survival rates than intermediate-1 and intermediate2–risk patients. At 182 days, the survival rates were 95.9%, 94.5%, and 88.5% for intermediate1–risk, intermediate-2–risk, and high-risk patients, respectively (Table S8). At 364 days, the survival rates were 93.6%, 91.6%, and 79.8%, respectively. At 728 days, the survival rates were 90.1%, 82.8%, and 66.8%, respectively, and at 1092 days, the rates were 86.5%, 78.7%, and 55.1%, respectively. Survival rates were generally lower in patients with a platelet count of < 100,000/mm³ at treatment initiation (81.4%, 70.6%, 58.7%, and 48.4% at Days 182, 364, 728, and 1092, respectively) than in those with platelet counts of ≥ 100,000/mm³ (95.5%, 91.8%, 84.2%, and 78.1% at Days 182, 364, 728, and 1092, respectively) (Table S9). Similarly, survival rates were lower in patients with baseline hemoglobin of < 10 g/dL (89.4%, 83.0%, 72.7%, and 64.0% at Days 182, 364, 728, and 1092, respectively) than in those with levels of ≥10 g/dL (98.6%, 95.8%, 90.5%, and 85.1% at Days 182, 364, 728, and 1092, respectively) (Table S10).

Figure 2.

Kaplan – Meier curve for overall survival. (A) Overall survival for all patients. (B) Overall survival by IPSS risk category. (C) Overall survival by platelet level. (D). Overall survival by hemoglobin level.

Abbreviations: Efficacy analysis set. CI, confidence interval; IPSS, International Prognostic Scoring System.

4. Discussion

This PMS of ruxolitinib treatment in patients with MF in Japan, which included all patients treated with ruxolitinib since its approval in 2014, provides real-world evidence for the safety and effectiveness of this drug in Japan. In this analysis, ruxolitinib showed improvements in spleen size and symptom burden, and it was generally well-tolerated by Japanese patients.

Ruxolitinib dosage for the treatment of MF ranges from 10 to 50 mg daily, although the dose can be adjusted according to the patient’s condition [8]. Data from this PMS suggest that for most patients ruxolitinib was administered within the dosage specifications, and that the average duration of ruxolitinib treatment was > 1.5 years.

Approval of ruxolitinib for the treatment of MF in Japan was mainly based on data from two Phase 3 clinical trials, namely COMFORT-I [9] and COMFORT-II [10], and an Asian multinational Phase 2, open-label, single-arm trial (Study A2202) [15]. The baseline characteristics of patients in this study largely reflected those of the general population with MF, with similar characteristics to those of patients who participated in the COMFORT studies [9,10] and Study A2202 [15]. Upon enrollment into this PMS, a higher proportion of patients had a hemoglobin level of < 10 g/dL (66.5% vs 56.7% and 45% in Study A2202 [15] and COMFORT-II [10], respectively). At baseline, a larger proportion of patients in Study A2202 had primary MF (67%) [15] compared with that in this PMS (46%) and the COMFORT studies (45% and 53% in COMFORT-I and COMFORT-II, respectively [9,10]).

While direct comparisons between the aforementioned studies and the PMS reported here cannot be made, the incidence of ADRs in this study (67.7%) was lower than that of 91.7% in Study A2202 and that of 81.1% in the pooled analysis of COMFORT-I and COMFORT-II (data on file). Similar to Study A2202, the incidence of ADRs in this PMS did not tend to increase in association with prolongation of the treatment duration [15].

No new or unexpected AEs with ruxolitinib treatment were observed in this PMS, and the findings were consistent with toxicities observed in the aforementioned studies. As a JAK1/2 inhibitor, ruxolitinib is known to inhibit JAK2, which regulates thrombopoietin and erythropoietin signaling [17,18]. Therefore, it is not surprising that cytopenias, especially anemia and decreased platelet count, is the most common AEs in patients with MF [9,10]. Consistent with these previous observations, anemia (27.8%) and decreased platelet count (21.4%) were the most frequent ADRs in this PMS. In this PMS, anemia and decreased platelet count rarely led to ruxolitinib discontinuation (3.4% and 3.3% of patients discontinued, respectively), whereas in Study A2202, these events led to dose reductions or interruptions more frequently among Japanese patients [15]. This suggests that these ADRs were effectively managed in most patients, possibly through dose modifications in the real-world setting.

Non-melanoma skin cancers (NMSCs), including basal cell, squamous cell, and Merkel cell carcinoma, have been reported in patients treated with ruxolitinib, however a causal relationship has not been established [8]. The incidence of NMSC in this PMS was lower (basal cell carcinoma, 0.1%; squamous cell carcinoma, 0.2%) compared with the rate reported in the phase 3b expanded-access study, JUMP (all NMSC, 2.7%) in which the majority of patients were European [19]. Most patients with MF where NMSCs have been reported have histories of extended treatment with hydroxyurea and prior NMSC or pre-malignancy skin lesions [20] and periodic skin examination is recommended for patients receiving ruxolitinib who are at increased risk for skin cancer [8].

The effectiveness of ruxolitinib in this PMS was evaluated by reduction in spleen size, the improvement of the PGA, and overall survival. Spleen response was achieved in 26.2% of patients at 6 months, and the mean reduction from baseline in palpable spleen size was 32%, indicating the effectiveness of ruxolitinib. This is consistent with the proportion of patients (30%) who experienced a ≥ 50% reduction in palpable spleen size at 24 weeks in a Japanese ruxolitinib early-access program, with a mean percentage decrease of 40.1% [21]. Similarly, in the Japanese subset of Study A2202, median percentage change in palpable spleen size ranged from − 29.6% to − 33.3% between Weeks 12 and 48 [22].

In addition to improving splenomegaly, ruxolitinib significantly reduces symptom burden. However, making direct comparisons concerning symptoms is challenging owing to the use of various methods for evaluating symptom responses. In Study A2202, 49% of the total study population [15] and 56% of Japanese patients achieved a ≥ 50% reduction from baseline in MF Symptom Assessment Form total symptom score at Week 24 [22]. These results were comparable to those observed in COMFORT-I (46%) [9]. In this PMS, patients’ symptom relief level was evaluated through PGA, and effectiveness was shown, with approximately half of the patients assessed as showing remarkable improvement or symptom improvement.

In the analysis of PGA response rate by patient factors, the adjusted odds ratios of two factors (disease duration and the presence or absence of concomitant medication for the primary disease) did not include 1, and no consistent trend was observed with the duration of disease. Patients on concomitant medication for the primary disease had a lower response, suggesting that these patients might have a more severe disease profile, which makes the treatment more challenging.

A pooled analysis of data from the COMFORT-I and COMFORT-II studies showed that ruxolitinib is associated with prolonged survival. At Week 144 (Day 1008), the Kaplan – Meier estimated overall survival rate was 78% in the ruxolitinib arm, 61% in the intent-to-treat control arm, and 31% in the crossover-adjusted control arm [23]. We found similar results in this study, with the overall survival point estimate being 71.8% at Day 1092, reflecting real-world prognoses in the ruxolitinib era. A nationwide survey of patients with primary myelofibrosis in Japan, which collected data for the period 1999–2015, reported a median survival duration of 47 months and a 3-year overall survival rate of 59% [5]. The study reported a lower overall survival rate, perhaps due to the advanced disease in the enrolled patients. However, this nationwide survey result cannot be directly compared with the results of our study and should therefore be interpreted with caution; we observed a higher overall survival rate in our study, possibly attributed to ruxolitinib use in the patient population investigated in this surveillance. Three factors – IPSS risk category, platelet count treatment initiation, and hemoglobin level at treatment initiation – showed different trends in the Kaplan – Meier curves by patient factor. Patients in the high-risk IPSS category, or those who began treatment with a platelet count of less than 100,000/mm³, or a hemoglobin level of less than 10 g/dL, generally had lower survival rates. Since these factors have been reported to be poor prognostic factors per IPSS or DIPSS-plus, the results are considered to reflect the subject background [24]. Several pooled analyses of COMFORT data have shown a survival advantage with ruxolitinib treatment compared with control treatment. This benefit was observed regardless of the patients’ anemia status at baseline, whether anemia developed during ruxolitinib treatment, or their transfusion status at Week 24 [11,25,26]. Consistent with these clinical trial results, a real-world study showed that lower hemoglobin levels at 6 months after starting ruxolitinib did not affect overall survival [27]. Furthermore, pooled analysis of the COMFORT data showed improved clinical outcomes for patients who initiated ruxolitinib earlier (≤12 vs > 12 months from diagnosis), including fewer cytopenia events, better spleen response, and longer OS [28] highlighting the importance of early intervention before cytopenic progression.

4.1. Limitations

The limitations of this study include its observational design, which lacks a control group, which makes it difficult to definitively associate the effects observed with the drug treatment. Another significant limitation lies in the different methods used to measure spleen shrinkage between this PMS and clinical trials. Unlike clinical trials that typically employ advanced imaging techniques such as magnetic resonance imaging or computed tomography to gauge spleen volume, this PMS used palpation – a much simpler method – for measuring spleen length. Therefore, caution is advised when making direct comparisons between these results and those obtained from clinical trials. Moreover, because the study was conducted under GPSP ordinance in Japan, source data verification was not carried out, as it is not a mandatory requirement under these regulations. The study does not provide detailed data on the transfusion status of patients before and after ruxolitinib treatment in the real-world setting in Japan. This represents a notable limitation of the study, particularly given the clinical significance of achieving transfusion independence after initiating ruxolitinib, which is often considered an indicator of successful therapy.

5. Conclusions

This post-marketing surveillance was conducted in more than 800 patients with MF in Japan, and the study demonstrated a certain degree of treatment effectiveness associated with ruxolitinib, with no new safety concerns. The data from this report align with those observed in reported clinical trials. This long-term assessment presents the ruxolitinib treatment data in patients with diverse baseline characteristics in a real-world clinical setting, and it would guide the hematologists in Japan in therapeutic decision-making for the benefit of the patients.

Funding Statement

This study was sponsored and funded by Novartis Pharmaceuticals.

Article highlights

• Ruxolitinib was the first JAK 1/2 inhibitor approved for the treatment of myelofibrosis (MF).

• In July 2014, the Japanese Ministry of Health, Labour and Welfare approved ruxolitinib for the treatment of MF. A condition of the approval was the implementation of an all-case surveillance on all patients treated with ruxolitinib until sufficient post-marketing data could be gathered.

• The post-marketing surveillance entailed a specified drug use results survey conducted on all ruxolitinib-treated patients starting from July 2014, with an observation period lasting for up to 3 years from the beginning of the treatment.

• A total of 892 patients were evaluated for safety and 891 for effectiveness.

• The most common adverse drug reactions were anemia and decreased platelet count. There were no new or unexpected safety signals in this study.

• Six months after treatment, this study showed that ruxolitinib reduced MFrelated splenomegaly in 26.2% of patients. Symptom improvement was observed in 52.0% of patients.

• These data demonstrated the effectiveness and safety of ruxolitinib therapy in realworld settings in Japan.

Author contributions

Yusuke Aruga, Wataru Hongo, and Weizhe Lu provided substantial contribution to the study design and/or collection, analysis, and/or interpretation of data; were involved in the drafting and/or critical review of the manuscript; provided final approval of the version to be published; and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Disclosure statement

All authors are employees of Novartis Pharma K.K. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Medical writing support was provided by Helen Findlow of Novartis Pharmaceuticals UK Ltd, London, UK, and funded by Novartis Pharmaceuticals.

Reviewer disclosure

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Ethical declaration

This was a prospective analysis of data from the post-marketing surveillance study of ruxolitinib in Japan. Since all treatments were at the physician’s discretion in actual clinical practice and the analysis was conducted anonymously, no patient consent was required for the present study under the regulations of Good Post-Marketing Study Practice. The study was conducted according to the regulations of Good Post-Marketing Study Practice.

Data availability statement

Novartis is committed to sharing access to anonymized patient-level data and clinical study reports from eligible studies with qualified external researchers. All data provided are anonymized to respect the privacy of patients who have participated in the trial in line with applicable laws and regulations.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/14796694.2025.2550924