Digital Features Digital Features for this article can be found at https://doi.org/10.6084/m9.figshare.20154776.
FormalPara Key Points
Atopic dermatitis (AD) has a multitude of symptoms that affect patients beyond the efficacy endpoints used by regulatory authorities when evaluating treatment success. To fully understand the ability of treatments to control disease, other endpoints and timepoints need to be considered.
To provide a better understanding of the impact of tralokinumab over time on multiple disease domains, we assessed the timing, magnitude, and sustainability of the effect over the entire 32-week treatment period of the ECZTRA 3 trial, including all patients initiated on tralokinumab irrespective of the response achieved at Week 16. Tralokinumab plus topical corticosteroids provided progressive and sustained improvements in both clinician- and patient-reported outcomes, including AD signs, symptoms, and health-related quality of life over 32 weeks.
These results provide a more complete overview of the benefits of tralokinumab over time, which may help inform clinical decisions.

Introduction

There remains a need for improved symptom control and reduced burden in patients with moderate-to-severe atopic dermatitis (AD). A large cross-sectional survey of US physicians and their patients with AD found that 42% of patients were identified as having inadequately controlled disease [1]. In recent years, advances in our understanding of the underlying pathophysiology have led to the discovery and regulatory approval of multiple new systemic treatments for AD [2].

AD is a chronic disease, requiring long-term management. The use of Investigator’s Global Assessment (IGA) 0/1 (clear/almost clear) and 75% improvement in Eczema Area and Severity Index (EASI-75) after 16 weeks of treatment as primary outcomes in clinical studies is largely driven by guidance from regulatory authorities such as the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA) [3]. However, these measures and timepoints do not sufficiently capture disease control, and patients with AD report a wide variety of therapeutic needs beyond lesion control, including itch and sleep disturbances [4]. A more comprehensive assessment of the impact on multiple disease domains and at timepoints beyond Week 16 is needed to inform clinical decisions to initiate, continue, discontinue, or modify treatment [5].

Tralokinumab is a fully human immunoglobulin G4 monoclonal antibody that specifically binds to the interleukin (IL)-13 cytokine with high affinity, preventing interaction with the IL-13 receptor and subsequent downstream IL-13 signaling [6,7,8]. Tralokinumab met all primary and secondary endpoints at Week 16 in the pivotal phase III trials (ECZTRA 1, 2, and 3), which included 1,976 adult patients with moderate-to-severe AD [9, 10]. Tralokinumab was recently approved for the treatment of moderate-to-severe AD in adult patients who are candidates for systemic therapy in the European Union, Canada, and the USA [11,12,13].

The ECZTRA 3 trial (NCT03363854) assessed the efficacy and safety of tralokinumab in combination with a topical corticosteroid (TCS) used on active lesions. Significantly more patients treated with tralokinumab every 2 weeks (q2w) plus TCS versus patients treated with placebo plus TCS achieved the primary endpoints of IGA 0/1 (38.9% vs. 26.2%; = 0.015) and EASI-75 (56.0% vs. 35.7%; < 0.001), as well as EASI-50 (79.4% vs. 57.9%; < 0.001) and EASI-90 (32.9% vs. 21.4%; = 0.022) at Week 16, as reported previously [10]. All confirmatory endpoints in the test hierarchy were also met. Most patients who achieved the primary endpoint maintained their response up to Week 32 on tralokinumab q2w plus TCS (89.6% maintained IGA 0/1 and 92.5% maintained EASI-75) and on tralokinumab every 4 weeks (q4w) plus TCS (77.6% maintained IGA 0/1 and 90.8% maintained EASI-75). To obtain further insight into the impact of tralokinumab on multiple disease domains and over time, here we assessed the timing, magnitude, and sustainability of the effect on signs, symptoms, and health-related quality of life (QoL) over the entire 32-week treatment period. To reflect clinical practice, we pooled all patients who started on tralokinumab in the initial treatment period irrespective of the response achieved at Week 16. Safety outcomes for ECZTRA 3 were previously reported for the different treatment arms [10].

Methods

Study Design and Patients

The design and methodology of ECZTRA 3 (NCT03363854) have been published previously [10]. Briefly, ECZTRA 3 was a double-blind, randomized, placebo-controlled, 32-week, phase III trial conducted across 68 sites in Europe and North America. Eligible patients were ≥ 18 years of age, with a diagnosis of AD for ≥ 1 year, an EASI score of ≥ 16, an IGA score of 3 or 4, and worst daily pruritus numeric rating scale (NRS) average score of ≥ 4.

Following a 2- to 6-week screening period, including washout of prior AD medications (2 weeks for topical treatments; 4 weeks for systemic treatments), patients were randomized 2:1 to subcutaneous tralokinumab 300 mg q2w with TCS as needed or placebo q2w with TCS as needed for 16 weeks, and received tralokinumab 600 mg (loading dose) or placebo on Day 0 (Fig. 1). At Week 16, patients who achieved the clinical response criteria with tralokinumab (IGA 0/1 and/or EASI-75) were re-randomized 1:1 to tralokinumab q2w or q4w plus TCS as needed for a further 16 weeks; the objective of the continuation period at Weeks 16–32 was to evaluate the ability to maintain the Week 16 response (IGA 0/1 and/or EASI-75) with two different dosing interval options: q2w and q4w. Patients who achieved clinical response with placebo continued to receive placebo (q2w) to maintain blinding of the study, and patients not achieving the clinical response criteria at Week 16 (from tralokinumab or placebo) received tralokinumab q2w plus TCS as needed from Week 16 onwards (Fig. 1).

Fig. 1
figure 1

Trial design. AD atopic dermatitis, q2w every 2 weeks, q4w every 4 weeks, TCS topical corticosteroid

Patients were provided TCS (mometasone furoate 0.1% cream; Europe class 3 (potent); US class 4 (mid-strength)) in kit sizes of 180–200 g q2w, free of charge, and were instructed to apply a thin layer of supplied TCS once daily to areas with active lesions as needed. The quantity of TCS used was measured based on used and unused tubes returned at bi-weekly visits. Patients were instructed to apply an emollient twice daily to lesional skin only when TCS was not applied. Rescue treatment (topical and systemic medications) was permitted to control intolerable AD symptoms at investigator discretion.

Endpoints

Endpoints evaluating AD extent and severity up to Week 32 included the proportion of patients achieving IGA 0/1; the proportion of patients achieving 50%, 75%, and 90% improvement in EASI (EASI-50, EASI-75, and EASI-90); and percentage improvement in EASI versus baseline. Endpoints evaluating patient-reported outcomes up to Week 32 included relative improvement from baseline in weekly average of worst daily pruritus NRS, eczema-related sleep interference NRS, and Dermatology Life Quality Index (DLQI). The proportion of patients achieving a response equivalent to established levels of clinically meaningful improvements at Week 16 in three disease-specific domains—AD signs (EASI-50) [14], symptoms (≥ 3-point pruritus NRS improvement) [15], and QoL (≥ 4-point DLQI improvement) [16]—was also assessed. Other endpoints included quantity of concomitant TCS used up to Week 32 (assuming no TCS was used from non-returned tubes).

Statistical Analysis

Statistical analyses followed pre-specifications [10]. Post hoc analyses were based on the full analysis set and were conducted by pooling all patients treated with tralokinumab in the initial treatment period (n = 252) irrespective of the response achieved at Week 16 and the dosing regimen (q2w or q4w) received beyond Week 16 (Fig. S1, see Electronic Supplementary Material (ESM)). Additionally, pre-specified analysis by Week 16 response is provided. For binary endpoints, the differences in response rates between treatment groups were analyzed using the Cochran-Mantel-Haenszel test stratified by region and baseline disease severity. Patients with missing data or who received rescue medication prior to the visit were assumed to be non-responders. For continuous endpoints, changes from baseline were analyzed using a repeated-measurements model with an unstructured (compound symmetric when needed for convergence) covariance matrix among visits within patients. Data collected after initiation of rescue medication or permanent discontinuation of investigational medicinal product were excluded from the analysis, but each patient contributed information to all timepoints via the covariance matrix. For waterfall plots illustrating percentage improvement in EASI at the individual patient level, last observation carried forward was applied for patients using rescue treatment, for patients discontinuing study medication, and for patients with missing data for other reasons up to Week 32.

Results

Patients

Overall, 380 patients were randomized in ECZTRA 3 to receive tralokinumab q2w plus TCS (n = 253) or placebo q2w plus TCS (n = 127) in the initial treatment period, with 233 patients continuing to receive tralokinumab (q2w (n = 164) or q4w (n = 69)) during the continuation period (Weeks 16–32) (Fig. S1, see ESM). One patient from each treatment group did not receive a treatment dose and was excluded from the analysis. Baseline demographics and disease characteristics were similar across treatment groups (Table 1). Overall, 46% of patients had severe disease (IGA score of 4) at baseline with a mean EASI score of 29.4, a mean DLQI score of 17.5, and a mean weekly average of worst daily pruritus NRS score of 7.7.

Table 1 Patient baseline demographics and disease characteristics

Clinician- and Patient-Reported Outcomes up to Week 32

Lesion Extent and Severity

Analyzing the efficacy of tralokinumab treatment over the entire 32-week treatment period, irrespective of the response achieved at Week 16 and the dosing regimen (q2w or q4w) received in the continuation treatment period (n = 252), EASI-50 response rates were sustained from Week 16 (79.4%) to Week 32 (81.0%) in the pooled tralokinumab group, while EASI-75 and EASI-90 response rates continued to improve beyond Week 16 to 70.2% and 50.4%, respectively, at Week 32 (Fig. 2 and Fig. S2, see ESM).

Fig. 2
figure 2

EASI response rates at Weeks 16 and 32. Composite estimand (primary analysis): Patients who received rescue medication were considered non-responders. Patients with missing data were imputed as non-responders. EASI Eczema Area and Severity Index, EASI-50 at least 50% improvement in EASI, EASI-75 at least 75% improvement in EASI, EASI-90 at least 90% improvement in EASI, q2w every 2 weeks, q4w every 4 weeks, TCS topical corticosteroid

At Week 16, the least squares mean percentage improvement in EASI from baseline was 71.2% with tralokinumab (< 0.001 vs. placebo); this increased to 84.2% at Week 32 with continued tralokinumab (Table 2 and Fig. S3, see ESM). Figure 3 provides further granularity into EASI responses over time, including the outcome for each of the 252 patients initiated on tralokinumab plus TCS at Weeks 4, 16, and 32.

Table 2 Efficacy endpoints at Weeks 16 and 32
Fig. 3
figure 3

Distribution of percentage improvement in EASI from baseline in all patients initiated on tralokinumab q2w at a Week 4, b Week 16, and c Week 32. a, b Tralokinumab q2w + TCS (n = 252), full analysis set. LOCF applied for patients using rescue medication/discontinuing study medication during the initial period (last observation available prior to using rescue medication/discontinuing study medication during the initial period). c Tralokinumab q2w/q4w + TCS (n = 252), full analysis set. LOCF applied for patients using rescue medication (during Weeks 16–32) or discontinuing study medication (last observation available prior to using rescue medication (during Weeks 16–32) or discontinuing study medication) and for patients with missing data for other reasons at Week 32 (last observation available). EASI Eczema Area and Severity Index, EASI-50 at least 50% improvement in EASI, EASI-75 at least 75% improvement in EASI, EASI-90 at least 90% improvement in EASI, LOCF last observation carried forward, q2w every 2 weeks, q4w every 4 weeks, TCS topical corticosteroids

Video 1 Percentage improvement in EASI from baseline to Week 32 in all patients initiated on tralokinumab q2w. (MP4 5305 KB)

In the subgroup of patients who achieved the pre-defined response criteria (IGA 0/1 and/or EASI-75) at Week 16 and were subsequently re-randomized to tralokinumab q2w (n = 69) or q4w (n = 69) plus TCS, the proportion of patients achieving EASI-90 continued to increase to 72.5% and 63.8%, respectively, at Week 32 (Fig. S4a, see ESM). In the subgroup of patients who did not achieve the pre-defined response criteria at Week 16 and continued receiving tralokinumab q2w plus TCS (n = 95), 34.7% achieved EASI-90 at Week 32 (Fig. S4a, see ESM).

Patient-Reported Outcomes

Patient-reported outcomes, including pruritus NRS, sleep NRS, and DLQI, improved from baseline to Week 16 and were sustained through Week 32 in the pooled tralokinumab-treated patient group (n = 252; Table 2 and Fig. 4). Weekly average of worst daily pruritus NRS and eczema-related sleep interference NRS improved throughout the initial 16-week treatment period, with a mean improvement of 52.6% with tralokinumab plus TCS versus 38.4% with placebo plus TCS at Week 16 for itch, and an improvement of 63.3% with tralokinumab plus TCS versus 46.8% with placebo plus TCS for sleep (both < 0.001). Sustained improvements in pruritus NRS and sleep NRS of 59.4% and 70.8%, respectively, were observed through Week 32 in the pooled tralokinumab-treated patient group (Fig. 4a, b).

Fig. 4
figure 4

Least squares mean percentage improvement by visit in a weekly average of worst daily pruritus NRS, b weekly average in eczema-related sleep NRS, and c DLQI. Hypothetical estimand: treatments were reassigned at Week 16 and the placebo arm was only followed up to Week 16. The tralokinumab arm was followed beyond Week 16 and the different dosing (q2w or q4w) was ignored. Rescue medication was reset at Week 16. Data collected after permanent discontinuation of study medication or initiation of rescue medication were not included. In case of no post-baseline assessments before initiation of rescue medication, the Week 2 change was imputed as 0. Repeated-measurements model: Endpoint = Treatment*Week + Baseline*Week + Region + Baseline IGA. Compound symmetry was assumed for the covariance matrix. *p < 0.05 vs. placebo + TCS; **p < 0.01 vs. placebo + TCS; ***p < 0.001 vs. placebo + TCS. DLQI Dermatology Life Quality Index, IGA Investigator’s Global Assessment, NRS numeric rating scale, q2w every 2 weeks, q4w every 4 weeks, SE standard error, TCS topical corticosteroids

DLQI total scores improved with tralokinumab plus TCS versus placebo plus TCS, with a mean improvement of 65.4% with tralokinumab plus TCS compared to 49.1% with placebo plus TCS at Week 16 (< 0.001). In the pooled tralokinumab-treated group, improvements in DLQI were sustained with continued tralokinumab plus TCS treatment through Week 32 (Fig. 4). In addition, the improvement in DLQI from study baseline was sustained through Week 32 in Week 16 responder subgroups who were re-randomized to tralokinumab q2w (83%) and q4w (77%) during the continuation period (Fig. S4b, see ESM).

Evaluation of Clinically Meaningful Response at Week 16

More patients receiving tralokinumab plus TCS achieved a clinically meaningful response in each of the three disease-specific domains measured—AD signs (EASI-50), symptoms (pruritus NRS ≥ 3 improvement), and/or QoL (DLQI ≥ 4 improvement)—compared with placebo plus TCS at Week 16 (Table 3). Most patients (89.9%) who were initiated on tralokinumab q2w plus TCS achieved a clinically meaningful response in at least one of the three domains at Week 16, and 75.3% achieved a response in AD signs in combination with symptoms or QoL (≤ 0.001 vs. placebo for both) (Table 3 and Fig. 5). Of the patients who were initiated on tralokinumab q2w plus TCS, 53.4% (132/247) achieved a clinically meaningful improvement in all three domains at Week 16 (difference vs. placebo, 25.0%; 95% confidence interval 14.9–35.2; < 0.001).

Table 3 Composite efficacy endpoints at Week 16
Fig. 5
figure 5

Composite endpoint response rates at Week 16. DLQI Dermatology Life Quality Index, EASI Eczema Area and Severity Index, EASI-50 at least 50% improvement in EASI, NRS numeric rating scale, q2w every 2 weeks, TCS topical corticosteroids

Topical Corticosteroid Use up to Week 32

Tralokinumab plus TCS-treated patients used less of the supplied TCS compared with placebo plus TCS-treated patients at Weeks 15–16 (11.4 g (SE, 1.6) vs. 20.2 g (SE, 2.3); = 0.002) and TCS use remained consistently low with continued tralokinumab treatment, with mean TCS use during Weeks 16–32 ranging from 9.2 to 13.6 g (SE, 1.2–2.0) every 2 weeks (Fig. 6).

Fig. 6
figure 6

TCS use from baseline to Week 32 by visit. Assuming no TCS was used from non-returned tubes. Hypothetical estimand: treatments were reassigned at Week 16 and the placebo arm was only followed up to Week 16. The tralokinumab arm was followed beyond Week 16 and the different dosing (q2w or q4w) was ignored. Rescue medication was reset at Week 16. Data collected after permanent discontinuation of study medication or initiation of rescue medication were not included. In case of no post-baseline assessments before initiation of rescue medication, the Week 2 change was imputed as 0. Repeated-measurements model: Endpoint = Treatment*Week + Baseline*Week + Region + Baseline IGA. Compound symmetry was assumed for the covariance matrix. *p < 0.05 vs. placebo + TCS; **p < 0.01 vs. placebo + TCS; ***p < 0.001 vs. placebo + TCS. IGA Investigator’s Global Assessment, q2w every 2 weeks, q4w every 4 weeks, SE standard error, TCS topical corticosteroids

Discussion

The pivotal ECZTRA 3 trial was designed to evaluate the efficacy of tralokinumab plus TCS versus TCS alone at Week 16 according to pre-defined thresholds established by guidance from regulatory authorities. All primary and secondary objectives were met, confirming the efficacy of tralokinumab q2w plus TCS for EASI-75, IGA 0/1, pruritus NRS, SCORing Atopic Dermatitis, and DLQI observed at Week 16 [10]. Tralokinumab plus TCS was well tolerated, with an overall frequency and severity of adverse events comparable with TCS alone over 32 weeks [10].

Since AD is a chronic condition requiring long-term treatment, in this post hoc analysis we evaluated the efficacy of tralokinumab plus TCS as needed over 32 weeks. Data were pooled from all patients initiated on tralokinumab q2w, no matter the response achieved at Week 16 or the dosing regimen (q2w or q4w) used between Weeks 16 and 32. This provides a better understanding of the clinical response achieved over time in all patients initiated on tralokinumab treatment at study start, reflecting clinical practice not bound by the design limitations of a pivotal phase III trial. In addition, because no single outcome captures the full burden of a multidimensional disease such as AD, a more holistic evaluation of the Week 16 response was performed.

Importantly, this analysis demonstrates that additional patients achieved the clinical response targets preferred by regulatory authorities (IGA 0/1 or EASI-75) beyond Week 16, indicating that response rates progressively improve over time with continued tralokinumab therapy. The proportion of patients achieving EASI-75 increased from 56% at Week 16 to 70% at Week 32, and the proportion achieving EASI-90 progressively improved to 50% at Week 32 with no indication of reaching a plateau (Fig. S2c, see ESM). This is clinically relevant, as Week 16 appears to be too early for evaluating the full benefit of tralokinumab on lesions in some patients. However, this analysis also identified a group of “super-responders” achieving EASI-75, or even EASI-90 response, after only 4 weeks of treatment with tralokinumab q2w plus TCS (Fig. 3a; Video 1, online version only).

Improvements observed with tralokinumab plus TCS in disease domains important to patients with AD, such as itch, sleep, and DLQI, occurred early, within the first few weeks of therapy, and were then sustained through Week 32.

Examining the Week 16 response from a holistic perspective, the vast majority (89.9%) of patients treated with tralokinumab q2w plus TCS achieved a clinical response equivalent to established levels of clinically meaningful improvements in AD signs (EASI-50), symptoms (pruritus NRS improvement ≥ 3), or QoL (DLQI improvement ≥ 4), and 75.3% achieved a meaningful reduction in lesion extent and severity in combination with an improvement in either pruritus or QoL.

The range of outcomes reported in this study shows progressive and sustained improvements using tralokinumab in both clinician- and patient-reported outcomes. When evaluating treatment success in clinical practice for a chronic disease such as AD, it is important to look beyond the primary endpoints and timepoints recommended by regulatory authorities. Considering the chronicity of AD, the impact of an AD treatment must be considered against the short- and long-term needs of each patient and the relevant risks and benefits of existing or alternative treatment options [5, 17]. Including patients in shared decision-making maximizes patient adherence and treatment success [18].

Limitations of this analysis include post hoc pooling of patients after re-randomization to two different dosing regimens (q2w and q4w) at Week 16. As reported previously, both tralokinumab q2w and q4w treatment regimens were able to maintain an EASI-75 response at Week 32 in nine out of ten patients [10]. Furthermore, the improvements in EASI-90 response rates and DLQI were similar between the q2w and q4w treatment arms during Weeks 16–32 (Fig. S4, see ESM); this supports pooling of the two treatment arms in the current analysis and expected real-world clinical practice.

The high response rates at Week 16 in the group treated with TCS alone should also be acknowledged. This may reflect the provision of TCS (mometasone furoate 0.1% cream) free of charge at bi-weekly visits in ECZTRA 3. Importantly, all patients treated with tralokinumab used approximately 50% less TCS at Week 16 and the mean use remained low (around 5–7 g/week) in the continuation period (Fig. 6), further demonstrating the effectiveness of tralokinumab.

Another limitation is the lack of a placebo comparator arm beyond Week 16. When designing the ECZTRA 3 study, comparison between tralokinumab plus TCS versus TCS alone was limited to Week 16. In order to ensure that all patients with uncontrolled moderate-to-severe AD at Week 16 had access to active systemic therapy, patients initially randomized to placebo who did not achieve the primary outcomes at Week 16 were assigned to receive tralokinumab q2w in addition to TCS during Weeks 16–32. This prevented the inclusion of the majority of patients initiated on placebo in any comparative analyses beyond Week 16.

Despite these limitations, the current analyses provide a more complete picture of the benefits of tralokinumab over time, which may help guide clinicians in their decision-making. It is difficult to compare the results of tralokinumab and other systemic treatment options, mainly because no comparative study has been carried out and between-trial comparisons should be made with caution. Based on recent trials comparing Janus kinase (JAK) inhibitors and dupilumab, JAK inhibitors appear to have a faster onset of action than dupilumab [19,20,21]; however, the response rates seem to even out over time. The proportion of patients achieving EASI-75 at Week 32 with tralokinumab plus TCS reported in our analysis (70%) is similar to the proportion of patients achieving EASI-75 over time reported in other systemic-plus-TCS studies in adults with moderate-to-severe AD [22, 23].

Conclusions

In summary, tralokinumab plus TCS as needed provided progressive and sustained improvements over 32 weeks in the extent and severity of AD and in patient-reported outcomes in patients with moderate-to-severe AD. TCS use, as needed, remained low in patients treated with tralokinumab compared with placebo over 32 weeks, demonstrating the TCS-sparing effects of tralokinumab.