Clinical Report

Accepted Jan 23, 2018; Epub ahead of print Jan 24, 2018

Acta Derm Venereol

Corr: Gil Yosipovitch, MD, Department of Dermatology and Itch Center, University of Miami Hospital, 1600 NW 10th Avenue, Miami, FL 33136, USA. E-mail: yosipog@gmail.com

Atopic dermatitis (AD) is a chronic inflammatory skin disease that affects approximately 15–30% of children and 2–10% of adults; up to 90% of cases present as mild or moderate (1–3).Hallmarks of AD include pruritus and pruritus-induced scratching, which lead to exacerbation of symptoms and worsening of disease (4). Although the pathophysiology behind AD-associated pruritus is under investigation, it is believed to be a complex pathway that involves the interaction of neuropeptides with keratinocytes, immune cells and nerve fibres (5).

AD-associated pruritus often results in sleep disturbance, psychosocial morbidity and reduced quality of life (QoL) for patients and caregivers. Elevated stress levels and sleep problems can exacerbate pruritus pathways (4, 6, 7). The “itch-scratch cycle” describes the pattern of pruritus that leads to scratching, and the scratching that leads to further inflammation (8). Patients may be unaware they are scratching during sleep, and young patients may additionally experience sleep dysfunction, which can lead to emotional and psychological issues in adolescence (5, 8). Stress and anxiety have been associated with neuropeptides that can induce further pruritus and subsequent scratching (5). In addition, the damage to skin because of scratching can lead to skin infection and scarring and possibly contribute to the risk for ‘atopic march’, which is the progression to other atopic diseases, such as allergic rhinitis and asthma (5, 9, 10). Treatments that provide rapid relief and control of pruritus are important parts of therapy to reduce the occurrence of exacerbation of disease and improve sleep and QoL (4). Treatment guidelines from the American Academy of Dermatology (AAD) and the American Academy of Allergy, Asthma & Immunology (AAAAI)/American College of Allergy, Asthma & Immunology (ACAAI) do not provide specific guidance for addressing AD-associated symptoms such as pruritus (4, 11). However, it is important that therapy reduces pruritus quickly, regardless of disease or pruritus severity or other demographics or characteristics.

Both the AAD and AAAAI/ACAAI guidelines recommend topical pharmacological agents such as topical corticosteroids (TCSs) and topical calcineurin inhibitors (TCIs) to treat AD, while noting the potential of these agents to reduce the severity of pruritus (4, 11). Although TCIs and TCSs have been shown to reduce the severity of pruritus (5, 12, 13), they are associated with a range of adverse effects, such as burning and stinging (TCIs) and skin atrophy (TCSs), although skin atrophy is rare with appropriate use of low- or mid-potency TCSs (4, 11, 14). A boxed warning for TCIs indicates a theoretical risk for malignancy, and superpotent TCSs have been associated with lymphoma (11, 15). Consequently, additional topical non-steroidal medications are needed that provide rapid relief of pruritus in AD with minimal adverse effects.

Crisaborole ointment is a non-steroidal phosphodiesterase 4 (PDE4) inhibitor for the treatment of mild-to-moderate AD (16). Results of several early phase II trials evaluating crisaborole demonstrated an improvement in pruritus in patients with AD (17–19). Crisaborole was later evaluated in 2 large, identically designed, double-blind, vehicle-controlled, phase III studies (AD-301: NCT02118766; AD-302: NCT02118792). In these studies, patients ≥2 years old with mild or moderate AD were randomly assigned 2:1 to receive crisaborole or vehicle ointment twice daily for 28 days (20). Significantly more crisaborole-treated than vehicle-treated patients achieved the primary endpoint of success in Investigator’s Static Global Assessment [ISGA] score (AD-301: 32.8% vs 25.4%, p = 0.038; AD-302: 31.4% vs 18.0%, p < 0.001), where success was defined as a score of clear [0] or almost clear [1], with a ≥ 2-grade improvement from baseline at day 29 (20). Improvement in severity of pruritus at day 29 was a secondary endpoint. Significantly more crisaborole-treated patients than vehicle-treated patients experienced improvement in pruritus (pooled data: 63% vs 53%, p = 0.002), which was defined as achievement of a pruritus severity score of none [0] or mild [1], with a ≥ 1-grade improvement from baseline (20). Crisaborole was well-tolerated; application site pain (primarily burning or stinging) was the most common treatment-related adverse effect (crisaborole 4.4%; vehicle 1.2%) (20). Because the primary analysis of these studies evaluated relief of pruritus at the end of treatment, the objective of the post hoc analysis presented herein was to explore the impact of crisaborole treatment on achievement of early relief of pruritus from the pooled results of the 2 phase III studies.

The study design and methods of the 2 phase III studies have been published previously (20). Briefly, 2 multicentre, randomised, double-blind, vehicle-controlled phase III studies were performed to assess the efficacy and safety of crisaborole in patients ≥2 years old with mild-to-moderate AD. Patients were randomly assigned 2:1 to receive either crisaborole ointment 2% or vehicle ointment. Treatments were then applied twice daily for 28 days (Fig. 1). Study protocols were developed and conducted in accordance with the principles of Good Clinical Practice and local country-specific regulatory requirements. An institutional review board approved all study protocols, informed consent/assent forms and relevant supporting data at each investigational centre.

The primary endpoint of the studies was success in ISGA, which was defined as achieving a score of 0 or 1, with a 2-grade improvement from baseline at day 29, based on a 5-point scale of clear [0] to severe [4]. The secondary endpoint was the proportion of patients with an ISGA score of clear [0] or almost clear [1] at day 29.

Exploratory endpoints included pruritus and signs of AD, which were both assessed on 4-point scales of none [0] to severe [3]. Improvement in pruritus or signs of AD were defined as none [0] or mild [1], with a ≥1-grade improvement from baseline. Pruritus data were measured twice daily by the patient or parent/caregiver using an electronic diary. Additional endpoints such as QoL measures included the Children’s Dermatology Life Quality Index (CDLQI) (patients 2–15 years of age), and the Dermatology Life Quality Index (DLQI) (patients ≥16 years of age) (21, 22).

In this post hoc analysis, early improvement of pruritus was defined as experiencing improvement at day 6 using the morning assessment. Endpoints in the post hoc analysis include the proportion of patients who experienced early improvement of pruritus, the proportion who experienced pruritus symptom improvement at earliest assessment (day 2) and the percentage reduction in pruritus severity over the first 6 days of treatment. The likelihood of early improvement in pruritus based on baseline demographics and disease characteristics was also evaluated. QoL scores for CDLQI and DLQI used in the correlation with early improvement of pruritus were based on the established minimal clinically important difference (MCID) for each scale (CDLQI: ≥2.5-point change from baseline; DLQI: ≥3.3-point change from baseline) (23, 24). The MCID is the smallest change in QoL score that patients perceive as beneficial and that would result in a change in patient therapy, in the absence of troublesome side effects and excessive cost (25). The correlation between early improvement in pruritus and treatment outcomes such as improvement in ISGA, QoL measures, other signs of AD and higher sleep scores (as a component of CDLQI) at day 29 were examined in crisaborole-treated patients. This analysis also evaluated the proportion of patients who began the study with itch and became itch free (defined as achieving a pruritus severity score of 0 among patients with baseline pruritus severity of mild or worse [score ≥1]) on days 2, 6 and 29, and the proportion of patients who experienced early improvement in pruritus and maintained improvement at day 29.

Populations from both studies were pooled for this analysis. Pruritus and other baseline characteristics were evaluated by use of descriptive statistics. Differences in proportions between treatment groups for improvement in pruritus were compared using the Fisher exact test. Differences in proportions between treatment groups for patients who became itch free were compared using normal approximation to binomial proportions. Differences in percentage change were compared using analysis of variance with a factor of treatment. Odds ratios with corresponding confidence intervals and p-values were found from a logistic regression with a factor of treatment. The missing observations were not imputed, statistical significance was set at a 0.05 level, and nominal p-values were displayed.

The pooled population from both studies included 1,016 patients treated with crisaborole and 506 patients given vehicle ointment. No significant differences were observed in the key baseline demographics across treatment groups (Table I). No significant differences were observed across treatment groups for baseline disease severity or baseline severity of pruritus.

A significantly greater proportion of patients treated with crisaborole demonstrated improvement at earliest assessment (day 2; p = 0.013) and early improvement of pruritus at day 6 (p < 0.001) than patients given vehicle (Fig. 2a). Additionally, a significantly greater percentage reduction in pruritus severity was observed on days 2–6 in patients treated with crisaborole than in those given vehicle (Fig. 2b). Early pruritus relief also persisted in most patients. Among 474 patients treated with crisaborole who experienced early improvement in pruritus (day 6), 321 (67.7%) still reported improvement at day 29. Also, most of the vehicle-treated patients who experienced early relief of pruritus still reported improvement at day 29 (n = 103/163 [63.2%]). More crisaborole patients with pruritus severity of mild or greater at baseline became itch free over the course of the study. The proportion of patients with itch at baseline who became itch free over the course of the study generally increased in both study arms. No difference was observed at day 2; however, significantly more patients receiving crisaborole became itch free at day 6 and day 29 (Fig. 3).

Treatment with crisaborole increased the likelihood of early improvement in pruritus in patients, regardless of varying baseline disease characteristics and demographics. For example, early improvement in pruritus was more likely with crisaborole treatment regardless of baseline pruritus severity (mild, p = 0.008; moderate, p = 0.002; severe, p = 0.001) (Fig. 4a). Patients receiving crisaborole were also more likely to experience early improvement in pruritus regardless of their baseline ISGA (Fig. 4b) and percentage of affected body surface area (BSA) at baseline (Fig. 4c).

Patients receiving crisaborole were significantly more likely to achieve early improvement in pruritus among most age groups (p < 0.05), except for patients aged 12–17 years, though there was a trend towards early pruritus improvement with crisaborole in this age group (odds ratio: 1.502; 95% CI 0.931– 2.423; p = 0.096) (Fig. S1a). Crisaborole demonstrated a significant likelihood of a patient experiencing early improvement in pruritus regardless of sex (male, p = 0.009; female, p < 0.001) (Fig. S1b).

Among crisaborole-treated patients, early improvement in pruritus was more likely to be correlated with improvement in disease. For example, patients who experienced early improvement in pruritus were significantly more likely to have experienced treatment success (p < 0.0001) and an ISGA score of clear (0) or almost clear (1) (p < 0.0001) at day 29 than those who did experience early pruritus improvement (Fig. 5). Similarly, crisaborole-treated patients who experienced early improvement in pruritus were more likely to have experienced improvement in most signs of AD by day 29 with the exception of exudation (Fig. 6).

Correlations were also observed between early relief of pruritus and some QoL measures in crisaborole-treated patients. Crisaborole-treated patients experiencing early improvement in pruritus were significantly more likely to experience an MCID in CDLQI and DLQI scores at day 29 than those who did not see early improvement in pruritus (p < 0.0001 and p = 0.0497, respectively) (Fig. 7). Early improvement in pruritus in crisaborole-treated patients was also associated with better scores in certain QoL components, particularly sleep, which is a component of the CDLQI. More patients who had early improvement in pruritus while receiving crisaborole reported that their pruritus had affected their sleep “only a little” or “not at all” at day 29 than those who did not have early improvement in pruritus (odds ratio: 2.656; 95% CI 1.708–4.130; p < 0.0001).

In this post hoc analysis that evaluated the effect of crisaborole on early relief of pruritus, improvement in pruritus was observed in significantly more patients treated with crisaborole than in those given vehicle as early as day 2 of the study. Most crisaborole-treated patients also experienced improvement in pruritus at day 6, regardless of baseline pruritus severity. In addition, more crisaborole-treated patients who had pruritus at baseline became itch free, with more than 20% experiencing complete relief of pruritus by day 6. Crisaborole treatment was more likely to lead to early improvement in pruritus regardless of various baseline characteristics, including baseline pruritus severity, disease severity (by ISGA) and percentage of affected BSA.

Rapid relief of pruritus is necessary to break the itch-scratch cycle and associated atopic march and to prevent chronicity, learned defective sleep patterns and negative physical and psychological complications and sequelae (5, 26–28). Early and effective intervention is particularly important in preserving QoL in young patients with AD (28). In the current analysis, early improvement in pruritus after crisaborole treatment was more likely to be associated with improvement in disease, such as most signs of AD at day 29 and improvement in ISGA score. Crisaborole-treated patients seeing early improvement in pruritus were also more likely to have higher scores in QoL per DLQI and CDLQI at day 29 than those who did not experience early improvement in pruritus. Additionally, early improvement in pruritus among crisaborole-treated children was associated with better scores in sleep (based on a component of the CDLQI) at day 29 than in those who did not have early pruritus improvement. Although the mechanisms are complex and not fully elucidated, crisaborole is believed to relieve pruritus through inhibition of PDE4, which is involved in production of proinflammatory cytokines through degradation of cyclic adenosine monophosphate (cAMP) (29). Elevated cAMP-specific PDE activity in mononuclear lymphocytes has been reported in patients with AD (30). Use of PDE4 inhibitors has shown that increased intracellular cAMP might contribute to the reduction of inflammation and itch (31). However, the exact mechanism is not clear and warrants further study based on the rapid relief observed in this analysis.

In addition to understanding the causes of pruritus, it is important to assess its severity because it is the cardinal symptom of AD. Various scales have been used to assess pruritus, including the verbal rating scale (VRS), the visual analogue scale (VAS) and the numeric rating scale (NRS)  (32). Where the VRS uses a 4-point verbally administered scale from 0 to 3 with increasing intensity, the VAS and NRS are 11-point scales that grade severity from 0 to 10, increasing in intensity. An analysis comparing the use of these tests in chronic pruritus found that they all had high validity and concurrent validity in pruritus intensity assessment (32). The scale used to measure pruritus in this study was a 4-point scale administered via electronic diary (20). The mean percentage reduction from baseline at day 6 in crisaborole-treated patients in this analysis was 42%. Because different measuring scales were used in this study, it is difficult to compare itch reduction with that found in other studies because of differing definitions of pruritus severity and differences in how the scales were administered. In addition, a limitation of this analysis was that the pruritus severity scale had not been validated at the time of publication development.

Although it is difficult to directly compare the early reduction in pruritus after use of crisaborole with the reduction time with other topical agents such as TCSs and TCIs, this analysis shows that pruritus relief with crisaborole is at least comparable with what has been reported with those agents (12, 13, 33). It is also important to consider the highly tolerable safety profile of crisaborole. The most commonly reported adverse event from the phase III studies was application site pain (including skin burning or stinging), which was reported in 4.4% of crisaborole-treated patients (20). This is especially important to consider with younger patients because discomfort can lead to issues of adherence. Therefore, the combination of rapid relief of pruritus and its safety profile makes crisaborole a suitable non-steroidal topical therapy option for the management of mild-to-moderate AD.

In this post hoc analysis, a greater proportion of crisaborole-treated patients experienced early relief of pruritus. Crisaborole-treated patients were more likely to experience early relief of pruritus regardless of most baseline characteristics, and early relief of pruritus was associated with improved treatment outcome measures.

The authors thank the study patients, investigators and investigational sites, whose participation made these studies possible. We also thank Huaming Tan, PhD, of Pfizer Inc for performing some of the statistical analyses. Robert J. Schoen, PharmD, and Corey Mandel, PhD, of ApotheCom (Yardley/San Francisco) provided writing and editorial assistance, which was funded by Anacor Pharmaceuticals, Inc., a wholly owned subsidiary of Pfizer Inc.

This study was funded by Anacor Pharmaceuticals, Inc., a wholly owned subsidiary of Pfizer, Inc.

Conflicts of interest: GY conducted clinical trials or received honoraria for serving as a member of the scientific advisory boards of Opko, TREVI, Eli Lilly, Menlo, Regeneron, Sanofi, Galderma, Pfizer, Novartis and Sienna, and received research funds from Pfizer, Allergan, GSK/Stiefel and the Leo Foundation. LFSG has served as an investigator for Pfizer, Incyte, Leo and GSK; as an advisor for Pfizer, Medimetriks, Sanofi, Regeneron, Leo and GSK; and as a speaker for Pfizer, Sanofi-Aventis, Regeneron and Leo. ML received research funds from AbbVie, Boehringer Ingelheim, Celgene, Eli Lilly, Incyte, Janssen, Johnson & Johnson, Leo, Medimmune/AstraZeneca, Novartis, Pfizer, Scioderm, Valeant and ViDac; and has served as a consultant for Allergan and Dr. Reddy’s Laboratories. JS has served as an investigator for Celgene, Eli Lilly, Menlo, Realm, Regeneron, Sanofi and F. Hoffmann-La Roche; as a consultant for AbbVie, Anacor Pharmaceuticals, Inc., Eli Lilly, Galderma, GSK, Incyte, Kiniksa, Menlo, Leo, Pfizer, Realm, Regeneron and Sanofi; and as a speaker for Regeneron and Sanofi. AMT is an employee and shareholder of Pfizer. LZ was an employee of Anacor Pharmaceuticals, Inc. during the conduct of these trials.