A Maximum-Use Trial of Ruxolitinib Cream in Adolescents and Adults with Atopic Dermatitis

An open-label MUsT of adolescents and adults with AD was conducted at one site in Canada and six sites in the USA from 20 June to 26 December, 2019. Eligible patients were aged ≥ 12–65 years with AD for ≥ 2 years, an IGA score ≥ 2, and BSA involvement ≥ 25%. Patients with an unstable course of AD (as determined by the investigator), other types of eczema, immunocompromised status, or any serious illness or medical condition that could interfere with study conduct, interpretation of data, or patients’ well-being were excluded from the study. Use of AD drugs during the washout period before baseline (biologics, five half-lives or 12 weeks; systemic immunomodulating agents, 4 weeks; topicals, 1 week) and during the study was prohibited.

Enrolled patients continuously applied 1.5% ruxolitinib cream BID for 28 days on all AD lesions identified at baseline (except lesions on the scalp), regardless of the reduction in lesion size. Tubes were counted and weighed at each study visit to assess adherence. At day 28, patients with no safety concerns could continue to apply 1.5% ruxolitinib cream BID, but only to active lesions, for an additional 28 days (optional). All patients had follow-up assessments 30 days after the last application of ruxolitinib cream. If patients used moisturizers before enrollment, they could continue throughout the study.

This study was conducted in accordance with Good Clinical Practice guidelines and the Declaration of Helsinki. Written informed consent/assent was provided by all patients before enrollment. The protocols were approved by the relevant institutional review board or ethics committee at each study site.

The primary objective was to assess the safety and tolerability of ruxolitinib cream. Safety was evaluated by the frequency, duration, and severity of adverse events (AEs). Changes from baseline in physical examinations, vital signs, and laboratory data for hematology, serum chemistry, and urinalysis were also analyzed. Secondary endpoints included the assessment of plasma concentrations of ruxolitinib and pharmacokinetic (PK) parameters, including maximum concentration (C_max), time to C_max, and area under the curve from 0 to 12 h. Blood samples for PK assessments were collected on days 1 (pre-application, 1, 2, 4, and 12 h post-application), 15 (pre-application and 1 h post-application), and 28 (pre-application, 1, 2, 4, and 12 h post-application). Exploratory endpoints included the proportion of patients who achieved IGA treatment success (IGA score of 0/1 with a ≥ 2-grade improvement from baseline), proportion of patients with a ≥ 75% improvement in the Eczema Area and Severity Index (EASI-75), proportion of patients with a ≥ 4-point improvement in the itch Numerical Rating Scale (NRS4) score, total affected BSA, and EASI score. Efficacy endpoints were measured on days 15, 28, and 56. Mean change from baseline in daily itch NRS score through day 56 was also assessed.

The sample size was based on other PK studies with ruxolitinib cream and was not based on statistical power calculations. Enrollment of approximately 40 patients was to ensure ≥ 32 evaluable patients, ≥ 12 of whom were aged 12–17 years (with ≥ 6 aged 12–15 years). The sample size was based on feasibility, a common practice for this type of study, and was considered adequate to achieve the objectives of the study, including characterization of PK parameters after a single application and at steady state [13‐15]. Patient demographics, baseline characteristics, patient disposition, efficacy, and all safety analyses were conducted using the full analysis set (all patients enrolled in the study with one or more applications of ruxolitinib cream). Data were analyzed by descriptive statistics using SAS^® software (SAS Institute Inc., Cary, NC, USA; version 9.4). Sample size, frequency, and percentages were used for categorical measurements; sample size, mean, median, standard deviation, standard error of the mean, range, and interquartile range (IQR) were calculated for continuous measurements. Data for continuous measurements are reported at baseline and at each visit (measured values, change from baseline, and percentage change from baseline, if applicable). The PK-evaluable population consisted of all patients with one or more applications of ruxolitinib cream and one or more post-application plasma PK samples with measurable concentrations of ruxolitinib. Pharmacokinetic concentration data were analyzed using descriptive statistics. Patients were also stratified based on the total affected BSA at baseline (25% to < 40% and ≥ 40%) and/or patient age (12–15, 16–17, and ≥ 18 years). Pharmacokinetic parameters were calculated from serial PK concentration data on days 1 and 28 by a noncompartmental analysis (Phoenix^® WinNonlin^® version 8.2; Certara, Princeton, NJ, USA). The steady-state plasma concentration (C_ss) of ruxolitinib was calculated as the average of all plasma concentrations across visits on days 15 and 28.

Forty-one patients were treated, with a median (IQR) age of 17 (15–36) years; 51.2% were male, and 68.3% were white (Table 1). The median affected BSA at baseline was 31.2% (IQR, 28.6–44.0%; range, 25.0–90.0%), and the mean (standard deviation [SD]) baseline EASI score was 20.8 (9.2). One patient had a treatment-emergent AE (TEAE) that led to study discontinuation (increased liver function test before the first treatment application, not considered related to treatment). One other patient discontinued, with the recorded reason being “Other.” Two additional patients completed the first treatment period but did not enter the extension period (reasons unknown). Thirty-seven patients (90.2%) continued with therapy in the extension period (days 28–56). Per the protocol, all of these patients were considered to have completed the study, regardless of whether they completed 28 days of treatment during the extension period. The daily amount of ruxolitinib cream applied (mean [SD]) decreased from the first 28-day period (21.8 [13.1] g) to the extension period (9.0 [7.5] g), consistent with the treatment of only active lesions during the extension period.

Approximately one-third of patients (31.7%) experienced TEAEs, most of which were mild or moderate in severity (Table 2). One patient had a serious AE not on a treated area (abscess on the lower extremity [investigator determined likely due to a previous tendon surgery in that area approximately 3 months before baseline]), which resolved and was considered unrelated to treatment. No fatal TEAEs were reported. Treatment-emergent adverse events determined by the investigator to be possibly related to treatment were experienced by four patients (9.8%), for a total of six treatment-related TEAEs (all grade 1 [neutropenia, n = 1; aspartate aminotransferase increase, n = 2; alanine aminotransferase increase, n = 1] or grade 2 [dyspnea, n = 1; hemoglobin decrease, n = 1]). The patients with neutropenia and decreased hemoglobin both had plasma concentrations of ruxolitinib well below the ex vivo half-maximal inhibitory concentration (IC₅₀) for inhibition of thrombopoietin (TPO)-stimulated phosphorylation of STAT3 (281 nM) in human whole blood, a surrogate marker to evaluate the potential effects on bone marrow production of blood cells (i.e., platelets) [8]. Both patients also had low neutrophil or hemoglobin levels, respectively, at baseline (Table 2). One patient, noted above, discontinued from the study because of an AE (sponsor and medical monitor’s decision) not considered to be related to treatment. This patient had increased baseline levels in alanine aminotransferase (43 IU/L; grade 1), aspartate aminotransferase (117 IU/L; grade 2), and lactate dehydrogenase (318 IU/L, grade 1). Three patients (as described above) had AEs that resulted in treatment interruptions (grade 1 neutropenia considered by the investigator as related to treatment, grade 2 dyspnea considered related to treatment when applying to lesions on the face [nonsmoker with no history of asthma or hay fever], and grade 3 lower extremity abscess considered not related to treatment). Minimal changes (decreases) in mean hemoglobin within the normal range were observed, as well as transient increases in platelets at day 15 that returned to baseline levels by day 28 (Fig. 1).

The mean (SD) C_ss of ruxolitinib was 104 (309) nM with a geometric mean of 26.5 nM (geometric coefficient of variation, 321%) during the first 28-day period in which all lesions at baseline were treated (Table 3). The overall mean plasma concentration peaked at 4 h after topical application on day 1 (Fig. S1a of the ESM) at 241 nM, with lower peaks among patients treated for < 40% BSA (Fig. S1b of the ESM) and adolescents vs adults (Fig. S1c of the ESM). The overall median trough plasma concentration was stable during the first 28 days and was lower on day 56 (Fig. 2a), likely due to the treatment of only active lesions during the second 28-day period, which had decreased substantially since baseline (Fig. 2b); concentrations were lower in patients treated for < 40% BSA (Fig. 2c), also likely due to fewer or a smaller area of treated lesions (Fig. 2d). The mean PK concentration–time profiles on day 28 were relatively flat (Fig. S1d of the ESM), with lower concentrations in patients treated for < 40% BSA (Fig. 1e). Overall, plasma concentrations were lower and the lesional area was smaller in adolescents vs adults (Fig. 2e, Fig. S1f of the ESM), consistent with the smaller treated area in adolescents (and less drug applied; Fig. 2f, Table 3). The peak plasma concentrations at steady state on day 28 were considerably lower than those observed after oral administration of ruxolitinib (15 mg BID) in healthy volunteers on day 10 in a phase I study [16] (Fig. 3). Modest peak/trough ratios were observed at day 28 (mean C_max/minimum concentration, 2.72; Table 3).

Five patients with the highest mean C_ss of ruxolitinib were further examined (Table S1 of the ESM). Of these patients, two patients with baseline affected BSA of 45% and 90% (average daily amount of 1.5% ruxolitinib applied during first 28 days, 45.5 g and 75.2 g, respectively; application rate, 2.5 and 2.2 mg/cm²) had C_ss exceeding the IC₅₀ for TPO-mediated pSTAT3. In these two patients, plasma concentrations were above the IC₅₀ for TPO-stimulated STAT3 phosphorylation on days 1, 15, and 28, and were below this threshold on day 56 (pre-application). No hematologic AEs were observed in these five patients with the highest mean C_ss of ruxolitinib during the full study period; one patient had a 57% reduction in platelets on day 28, but remained within the normal range throughout the study. Two of these patients reported no AEs during the study; one patient experienced grade 1 alanine aminotransferase (baseline, 18 IU/L; day 15, 94 IU/L) and aspartate aminotransferase elevations (baseline, 30 IU/L; day 15, 67 IU/L) that were considered related to treatment and resolved without dose modification; one patient (described above as a serious AE) had a grade 3 limb abscess considered unrelated to treatment; and one patient had peripheral edema and skin rash (both grade 1) that were considered unrelated to treatment and resolved without dose modification.

Mean (SD) systemic bioavailability was 2.5% (3.6%). There was no increase in C_max, concentration at 12 h post-application, and area under the curve from 0 to 12 h between day 1 and day 28, suggesting no ruxolitinib accumulation in plasma after repeated applications of ruxolitinib cream BID for 28 days (Table 3). There was no consistent pattern between change in hematologic laboratory parameters and plasma exposure except a transient increase in platelet count on day 15 (first study visit) that was plasma ruxolitinib concentration dependent. The mean platelet counts spontaneously decreased with continuous treatment and returned to (or near) baseline values by day 28 (the next visit).

The proportion of patients achieving IGA treatment success at day 15 was 20.0% and increased to 35.9% and 56.8% at days 28 and 56, respectively. Most patients had clear or almost clear skin (IGA score of 0/1) by day 28 (61.5%), with a further increase at day 56 (81.1%; Fig. S2 of the ESM); no patient had an IGA score of 0/1 at baseline. The mean (SD) total affected BSA at days 15, 28, and 56 was 15.8% (11.4%), 6.5% (8.2%), and 3.1% (5.4%), respectively; for context, the mean (SD) BSA at baseline was 38.1% (16.3%). The onset of itch relief was rapid, with a mean change from baseline in itch NRS, as measured through patient-reported daily diaries, of − 1.9 at day 1 (approximately 12 h after the first application; Fig. S3 of the ESM). Mean itch NRS continued to decrease; the mean change was − 5.3 at day 20, and reductions were maintained for the remainder of the study. The proportions of patients achieving NRS4 at days 15, 28, and 56 were 58.6%, 82.6%, and 90.5%, respectively. The mean (SD) EASI score at days 15, 28, and 56 was 6.3 (5.2), 2.4 (2.3), and 1.3 (2.2), respectively; for context, the mean (SD) score at baseline was 20.8 (9.2). The proportions of patients achieving EASI-75 at days 15, 28, and 56 were 30.0%, 79.5%, and 94.6%, respectively (Fig. S4 of the ESM). Representative clinical images are shown in Fig. S5 of the ESM.