Pruritus is a characteristic symptom of atopic dermatitis (AD),1 an inflammatory skin condition which affects up to a quarter of children and 5% of adults worldwide.2, 3 The itch–scratch cycle associated with pruritus causes distress to patients, reducing quality of life (QoL) and affecting sleep and daily activities.4-7 As AD is a chronic condition in which patients suffer from repeated phases of relapse and remission,1, 6 it is important to extend the periods of remission from pruritus and rash, in order to improve the quality of daily life.

The pathogenesis of allergic skin diseases is complex, and the definitive cause of pruritus in AD remains unclear, but cytokines appear to play a key role.8 In particular, interleukin (IL)-31 is a key mediator for pruritus in skin conditions including AD and prurigo nodularis,9-13 and appears to have proinflammatory and immunomodulatory functions as well as pruritogenic activity.14, 15

The humanized monoclonal antibody nemolizumab targets IL-31 receptor A,16 and in a recent 16-week, double-blind, phase III study, nemolizumab plus topical agents produced a greater reduction in pruritus associated with AD compared with placebo plus topical agents.17 The mean percentage change in pruritus visual analogue scale (VAS) score from baseline to week 16 favoured nemolizumab vs. placebo[difference −21·5%; 95% confidence intervals (CI) −30·2% to −12·7%; P < 0·001], and the mean percentage change in secondary endpoints such as the Eczema Area and Severity Index (EASI) confirmed the benefits of nemolizumab treatment (difference −12·6%; 95% CI −24·0% to −1·3%).17 Current Japanese guidelines for AD recommend the first-line use of topical agents, and oral antihistamines may be used as adjunctive therapy to reduce pruritus.18 Thus, by administering nemolizumab alongside current therapies, the design of this pivotal study reflected the management situation of many patients with AD and pruritus. A phase IIb study of nemolizumab administered concomitantly with topical corticosteroids (TCS) also reported sustained pruritus improvements in patients with AD over 24 weeks of treatment.19

Herein, we report data from two phase III clinical studies examining the effectiveness and safety of long-term (up to 68 weeks) nemolizumab, administered concomitantly with TCS and/or topical calcineurin inhibitors (TCI), in patients with AD with inadequately controlled moderate-to-severe pruritus.

Patients and methods Study design, treatments and blinding

We conducted two phase III, multicentre, long-term studies of nemolizumab for the treatment of pruritus associated with AD, which was inadequately controlled by topical agents and oral antihistamines. The study designs are shown in Figure S1 (see Supporting Information). Studies were conducted in compliance with the Declaration of Helsinki, Good Clinical Practice and Japanese regulatory ordinance. Trial documentation was approved by the institutional review boards at each centre. Patients (or their legal guardian) provided written informed consent prior to treatment.

Study-JP01 (JapicCTI-173740)

Patients were first enrolled into Part A (16 weeks, randomized, double-blind, placebo-controlled) and randomly assigned (2 : 1 ratio) to receive nemolizumab 60 mg or placebo every 4 weeks (Q4W) by subcutaneous injection (both plus TCS/TCI and/or oral antihistamines). Full details of Part A have been published.17 Patients completing Part A could enter a 52-week, open-label, long-term extension period (Part B); no additional selection criteria were imposed for Part B entry. All patients in Part B received nemolizumab 60 mg Q4W up to week 64, resulting in nemolizumab/nemolizumab and placebo/nemolizumab assessment groups.

Study-JP02 (JapicCTI-183894)

All patients received nemolizumab 60 mg Q4W up to week 48. Administration at baseline, week 4 and week 8 was by a medical professional. From week 12, half of the patients switched to self-injection.

In both studies, at the end of treatment, there was an 8-week follow-up period.

Patients

Eligible patients were aged ≥ 13 years, with a bodyweight of ≥ 30·0 kg, and a confirmed diagnosis of AD (as per the criteria of Hanifin and Rajka20) with pruritus. At the time of informed consent, patients were required to have a score of ≥ 3 on a five-level itch scale,21 indicating inadequate pruritic response, despite treatment with medium-potency (or stronger)6 TCS/TCI administered at a stable dose for ≥ 4 weeks, and oral antihistamines administered at a stable dose for ≥ 2 weeks, or an inability to receive such therapies. A VAS score for pruritus22 of ≥ 50 was also an inclusion criterion for both studies.

Exclusion criteria were any clinically relevant medical condition that could endanger the patient or render them inappropriate for study participation, abnormal laboratory values for liver enzymes or haematological parameters, or presence of diseases likely to affect the assessment of AD eczema and pruritus.

Prohibited concomitant therapies included antibody drugs (excluding nemolizumab), phototherapy and hyposensitization therapies, and systemic immunosuppressive treatments. Concomitant stable medium-potency TCS/TCI were used during Part A of Study-JP01, and TCS of any potency could be used during Part B of Study-JP01 and throughout Study-JP02.

Outcomes

The primary efficacy endpoint for Part A of Study-JP01 (the percent change in the weekly mean pruritus VAS score from baseline to week 16) has been described previously.17 For Part B of Study-JP01 and for Study-JP02, efficacy endpoints included the change over time in the following measures (where higher scores denote more severe symptoms): pruritus VAS score (range 0–100), five-level itch scale (range 0–4), the pruritus numeric rating scale (NRS, range 0–10),23 the EASI (range 0–72) score,24 the static Investigator’s Global Assessment (sIGA, range 0–5) score,25 and the Insomnia Severity Index (ISI, range 0–28).26 In addition, the total Dermatology Life Quality Index (DLQI, range 0–30),27 the Patient-Oriented Eczema Measure (POEM, range 0–28) score,28 and the mean daily quantity of topical agents used during the study period were assessed in Study-JP01. Patients used an electronic diary (daily from baseline to week 16, then weekly) to input pruritus VAS, NRS and five-level itch scale scores. EASI and sIGA were assessed by the investigator. The ISI, DLQI and POEM were completed by patients at study visits.

Additional efficacy measures were the proportions of patients in both studies who achieved the following: a 50%, 75% or 90% decrease in the pruritus VAS score or the EASI score from baseline, a score on the five-level itch scale of ≤ 1, a decrease of ≥ 4 points from baseline in the pruritus NRS score, a decrease of two or more levels in the sIGA score (i.e. final score of ≤ 1), a score of ≤ 7 on the ISI, and an improvement of ≥ 6 points on the ISI. In Study-JP01, the proportions of patients with a decrease of ≥ 4 points from baseline in the total DLQI score [considered to be the minimal clinically important difference (MCID)],29 and a decrease of ≥ 4 points from baseline (MCID)30-32 in the POEM total score were calculated.

Safety endpoints included treatment-emergent adverse events (TEAEs), serious TEAEs, TEAEs requiring discontinuation or interruption of study treatment, and TEAEs of special interest. Injection-related reactions were defined as adverse reactions which developed within 24 h after nemolizumab administration. Severity of TEAEs was determined by the investigator as mild (discomfort without limiting normal activities of daily living), moderate (discomfort disturbing or affecting activities of daily living) or severe (disturbing work or normal activities of daily living).

Statistical methods

For Study-JP01, the target sample size for Part A was 204 (nemolizumab 136, placebo 68) as per the POWER procedure (t-test) in SAS software (SAS Institute Inc., Cary, NC, USA);17 no additional power calculations were conducted for Part B, which included patients who completed Part A. For Study-JP02, the target sample size was 80; this was intended to ensure that enough patients completed long-term treatment (accounting for people who dropped out) to allow for sufficient data to evaluate long-term efficacy and safety.

The modified intention-to-treat population included all patients who met the inclusion/exclusion criteria, who received at least one dose of study treatment, and had data available for evaluation. For Study-JP01, baseline for this analysis was at the time of randomization to Part A; for Study-JP02, baseline was at study entry. Summary statistics were recorded at each timepoint. Missing data were not imputed for the continuous endpoints but were imputed as nonresponse for binary endpoints. The safety analysis set included all patients who received at least one dose of the study treatment. Integrated safety results for all nemolizumab-treated patients in both studies are provided.

Results Patients

In total, 215 patients were randomly assigned to treatment in Study-JP01 (nemolizumab/nemolizumab, n = 143; placebo/nemolizumab, n = 72), of whom 206 (n = 139 and n = 67, respectively) proceeded to Part B. In Study-JP02, 88 patients received nemolizumab treatment (of whom 44 switched to self-injection at week 12), as shown in Figure S2 (see Supporting Information). Completion rates were high in both studies.

Baseline demographic data are shown in Table 1. Overall, the populations of the two studies were comparable, with the exception that in Study-JP02, a higher percentage of patients had a sIGA score of 4 or more, compared with Study-JP01. Due to differing medication usage rules, TCS use varied between Study-JP01 and Study-JP02. Overall, around 60% of patients in both studies had an allergic disease at baseline.

Table 1. Demographics and other baseline characteristics (modified intention-to-treat population) Study-JP01 Study-JP02 Nem/nem (n = 143) Plb/nem (n = 72) Nem (n = 88) Male sex, n (%) 93 (65·0) 48 (66·7) 56 (63·6) Age (years), median (Q1–Q3) 39·0 (27·0–47·0) 40·5 (29·5–48·0) 40·0 (32·0–46·0) Disease duration (years), median (Q1–Q3) 30·3 (19·2–38·5) 28·9 (19·2–38·1) 31·0 (20·0–38·5) Pruritus VAS score, median (Q1–Q3)a 75·7 (69·0–82·1) 75·1 (69·1–82·1) 78·9 (70·9–87·6) Pruritus NRS score, median (Q1–Q3)a 7·3 (6·9–8·0) 7·4 (7·0–8·0) 7·7 (6·9–8·4) 5-level itch scale score, median (Q1–Q3)a 3·0 (3·0–3·1) 3·0 (3·0–3·0) 3·0 (3·0–3·2) EASI score, median (Q1–Q3) 24·2 (16·9–36·1) 22·7 (15·5–33·8) 27·0 (18·7–37·4) sIGA score of 4 or more, n (%) 61 (42·7) 27 (37·5) 55 (62·5) ISI score n = 142 n = 72 n = 87 Median (Q1–Q3) 12·0 (8·0–18·0) 12·0 (8·0–16·0) 11·0 (7·0–16·0) DLQI score n = 136 n = 69 – Median (Q1–Q3) 12·0 (8·0–16·0) 12·0 (8·0–14·0) – POEM score n = 142 n = 72 – Median (Q1–Q3) 22·0 (18·0–26·0) 20·5 (15·0–25·0) – Baseline treatment, n (%) Topical therapyb 143 (100·0) 72 (100·0) 87 (98·9) Potent/highly potent TCS 0 0 85 (96·6) Medium-potency TCS 139 (97·2) 70 (97·2) 31 (35·2) TCI 59 (41·3) 29 (40·3) 43 (48·9) Oral antihistamines 128 (89·5) 63 (87·5) 82 (93·2) Nonsedating 127 (88·8) 61 (84·7) 70 (79·5) Sedating 17 (11·9) 11 (15·3) 16 (18·2) Use of TCS/TCI (g daily), median (Q1–Q3)c 2·9 (1·6–5·7) 2·9 (1·9–4·8) – Allergic diseases at baseline, n (%) 94 (65·7) 45 (62·5) 52 (59·1) Seasonal allergy 37 (25·9) 23 (31·9) 18 (20·5) Rhinitis allergic 38 (26·6) 15 (20·8) 19 (21·6) Conjunctivitis allergic 32 (22·4) 13 (18·1) 15 (17·0) Food allergy 21 (14·7) 11 (15·3) 16 (18·2) Asthma 21 (14·7) 8 (11·1) 23 (26·1) All patients included in the trial (100%) were Japanese. DLQI, Dermatology Life Quality Index; EASI, Eczema Area and Severity Index; ISI, Insomnia Severity Index; Nem, nemolizumab; NRS, numeric rating scale; Plb, placebo; POEM, Patient-Oriented Eczema Measure; Q, quartile; sIGA, static Investigator’s Global Assessment; TCI, topical calcineurin inhibitors; TCS, topical corticosteroids; VAS, visual analogue scale. aThe pruritus VAS score, pruritus NRS score, and five-level itch scale score were the average measurement over the previous 24 h. bThe use of multiple agents was allowed. cThe median daily usage was calculated using data collected over a 4-week period. Efficacy outcomes

The percentage change in pruritus VAS score is shown in Figure 1a. In Study-JP01, the shift in the mean value of the pruritus VAS scores demonstrated a continuing trend towards reduced pruritus over time. In the nemolizumab/nemolizumab group, the decrease from baseline in pruritus VAS at week 68 was 65·9%. At the end of the 8-week follow-up period (12 weeks after the last administration), the pruritus VAS score showed only minimal increases, indicating that the effectiveness of nemolizumab against pruritus was durable. In Study-JP02, no differences were observed between the patients who self-injected or those who continued to receive administration from a medical professional (data not shown); overall, all patients receiving nemolizumab had a decrease in pruritus VAS scores from baseline at each study timepoint. The improvement in pruritus was similar in Study-JP01 and Study-JP02. A scatterplot indicating durable improvements in pruritus VAS over time and a graph showing the absolute VAS score are shown in Figure S3 and S4a (see Supporting Information).

Percentage change in (a) pruritus VAS scores and (b) EASI scores (modified intention-to-treat population). Fu1 and Fu2 denote 4 and 8 weeks after the end of the treatment period, respectively. Error bars denote standard error of the mean. Study-JP01 was double-blind until week 16 (denoted by shaded area).

BL, baseline; EASI, Eczema Area and Severity Index; Fu, follow-up; VAS, visual analogue scale.

Similar trends were observed in the change in EASI scores from baseline (Figure 1b); patients in the nemolizumab/nemolizumab group in Study-JP01 demonstrated a continued decrease in EASI during the long-term treatment period, and a decrease from baseline at week 68 of 78·2%. The reductions in EASI score were maintained after the end of treatment, with minimal changes during the follow-up period. The changes in EASI scores were comparable in Study-JP02. Absolute EASI scores are shown in Figure S4b; EASI scores were decreased to 5·6 (where 1·1–7·0 is defined as mild severity33) at week 68.

As shown in Figure 2, improvements in sleep (ISI) and DLQI were observed by week 16, and these improvements were maintained until the end of treatment. No rapid exacerbations were observed after the final administration of nemolizumab. The proportion of patients with a DLQI score ≤ 4 between baseline and the end of the follow-up period is shown in Figure S5 (see Supporting Information).

Proportion of patients with (a) ISI score of ≤ 7 and (b) a decrease of ≥ 4 points in DLQI score from baseline to the end of the follow-up period (modified intention-to-treat population). Fu2 denotes 8 weeks after the end of the treatment period. Study-JP01 was double-blind until week 16 (denoted by shaded area).

BL, baseline; DLQI, Dermatology Life Quality Index; Fu, follow-up; ISI, Insomnia Severity Iindex.

A continued decrease in TCS/TCI usage was observed in Study-JP01 (both groups) during the long-term administration period (Figure 3). Usage of TCS/TCI did not increase during the 8-week follow-up period (12 weeks after the last administration). Outcomes for all other efficacy endpoints are summarized in Table 2; in general, outcome measures showed a tendency towards improvement between weeks 16 and 68 (Study-JP01) and between weeks 16 and 52 (Study-JP02). A high proportion of patients (around 80%) achieved an improvement of ≥ 4 points (MCID) in POEM.

Daily usage of TCS and/or TCI from baseline to the end of the follow-up period (modified intention-to-treat population). The median daily usage was calculated using data collected over a 4-week period. Fu2 denotes 8 weeks after the end of the treatment period. Study-JP01 was double-blind until week 16 (denoted by shaded area).

Fu, follow-up; TCI, topical calcineurin inhibitors; TCS, topical corticosteroids.

Table 2. Summary of other efficacy endpoints (modified intention-to-treat population) Week 16 Week 68 (Study-JP01) or Week 52 (Study-JP02) Study-JP01 Study-JP02 Study-JP01 Study-JP02

Nem/nem

n = 143

Plb/nem

n = 72

Nem

n = 88

Nem/nem

n = 143

Plb/nem

n = 72

Nem

n = 88

Improvement in pruritus VAS, % 50% 34·3 13·9 45·5 57·3 58·3 55·7 75% 14·7 4·2 14·8 32·2 34·7 34·1 90% 6·3 2·8 4·5 14·0 13·9 20·5 5-level itch scale score ≤ 1, % 16·8 5·6 20·5 38·5 40·3 36·4 Improvement in pruritus NRS, % ≥ 4 points 32·2 12·5 44·3 49·7 56·9 55·7 Improvement in EASI, % 50% 51·7 40·3 56·8 79·7 75·0 80·7 75% 25·9 18·1 33·0 66·4 59·7 52·3 90% 7·0 4·2 12·5 40·6 33·3 27·3 Improvement in sIGA, % ≥ 2 points and a score of 0 or 1 5·6 5·6 8·0 28·7 16·7 12·5 Improvement in ISI, % n = 118 n = 61 n = 87 n = 118 n = 61 n = 87 ≥ 6 points 55·1 26·2 39·1 62·7 50·8 29·9 Improvement in POEM, %a n = 142 n = 72 – n = 142 n = 72 – ≥ 4 points 73·2 44·4 – 79·6 77·8 – Usage of TCS and TCI (g daily), median (Q1–Q3)a,b 2·42 (1·29–4·00) 2·91 (1·87–4·48) – 1·29 (0·55–2·64) 1·73 (0·82–2·96) – EASI, Eczema Area and Severity Index; ISI, Insomnia Severity Index; Nem, nemolizumab; NRS, numeric rating scale; Plb, placebo; POEM, Patient-Oriented Eczema Measure; Q, quartile; sIGA, static Investigator’s lobal Assessment; TCI, topical calcineurin inhibitors; TCS, topical corticosteroids; VAS, visual analogue scale. aStudy-JP01 only. bThe median daily usage was calculated using data collected over a 4-week period. Safety outcomes

Overall, TEAEs occurred in more than 90% of patients who received nemolizumab in the two studies (Table 3), but the majority were mild in severity; severe TEAEs occurred in < 5% of patients. The most common TEAEs were nasopharyngitis (33·9%) and AD (25·2%). The frequency of injection-related reactions decreased over time, to < 1% during long-term administration, with no occurrence of unexpected late-onset TEAEs (Table S1; see Supporting Information).

Table 3. Treatment-emergent adverse events (TEAEs) occurring after the first dose of nemolizumab (safety analysis set) Pooled nemolizumab (n = 298)a n (%) n per 100 PY Patients with ≥ 1 TEAE 281 (94·3) 75·7 Severe TEAEs 14 (4·7) 3·8 Moderate TEAEs 108 (36·2) 29·1 Mild TEAEs 266 (89·3) 71·6 Patients with ≥ 1 serious TEAE 28 (9·4) 7·5 Treatment modification due to TEAEs Discontinuation 14 (4·7) 3·8 Dose interruption 20 (6·7) 5·4 Injection-related reaction 22 (7·4) 5·9 Most frequently reported TEAEs (≥ 5% of patients in the pooled nemolizumab treatment group) by preferred term Nasopharyngitis 101 (33·9) 27·2 Atopic dermatitis 75 (25·2) 20·2 Blood creatine phosphokinase increased 27 (9·1) 7·3 Contact dermatitis 26 (8·7) 7·0 Influenza 26 (8·7) 7·0 Urticaria 24 (8·1) 6·5 Acne 22 (7·4) 5·9 Cellulitis 21 (7·0) 5·7 Headache 21 (7·0) 5·7 Dental caries 19 (6·4) 5·1 Upper respiratory tract inflammation 19 (6·4) 5·1 Gastroenteritis 17 (5·7) 4·6 PY, person-years. aIncludes all patients in the nemolizumab/nemolizumab group in Parts A and B of Study-JP01 (n = 143), all patients in the placebo/nemolizumab group who received nemolizumab during Part B of Study-JP01 (n = 67), and all of the patients who received nemolizumab in Study-JP02 (n = 88).

Cytokine abnormalities [increased level of thymus and activation-regulated chemokine (TARC)] were observed in 4·7% of patients (Figure S6; see Supporting Information). However, by 32 weeks after the start of treatment, TARC levels had returned to baseline, and were reduced still further after 56 weeks.

Discussion

In this analysis of data from two long-term (≥ 52 weeks) phase III studies of nemolizumab administered concomitantly with TCS/TCI, all of the measured efficacy outcomes were improved following initiation of nemolizumab, with effectiveness maintained or increased through the duration of the studies. Moreover, acute itchiness or flare of AD (e.g. relapse of pruritus, or worsening of the signs or extent of AD) were rarely observed during the 8-week follow-up period.

In patients enrolled in Parts A and B of Study-JP01, and who received nemolizumab for the entire 68-week treatment period, pruritus VAS decreased by 66% from the start of treatment. This compares with an improvement of 42·8%, which was previously reported at 16 weeks in Study-JP01.17 Although the absolute pruritus VAS score at baseline (74·9–78·4) indicated severe pruritus,22, 34 scores up to week 68 had decreased to a level (23·1–31·0) indicative of mild pruritus, suggesting a clinically meaningful improvement for patients. Although a higher percentage of patients in Study-JP02 had a sIGA score of 4 or more compared with Study-JP01, the degree of itchiness reached after long-term nemolizumab administration was the same in both studies.

Several immunotherapies have either recently been approved or are currently being developed for the tre