Anchored Matching-Adjusted Indirect Comparison of the Long-Term Maintenance of Efficacy of Tralokinumab and Lebrikizumab in Patients with Moderate-to-Severe Atopic Dermatitis

Abstract

Introduction

Atopic dermatitis (AD) is a chronic skin disease largely driven by interleukin (IL)-13, which is targeted by tralokinumab and lebrikizumab. Effective, long-term treatments are needed, and head-to-head studies comparing these available therapeutics have yet to be conducted. This analysis indirectly compared the maintenance of efficacy of tralokinumab and lebrikizumab at week 52 in week-16 responders.

Methods

An anchored matching-adjusted indirect comparison was performed using individual patient data (IPD) from the ECZTRA 1 and ECZTRA 2 tralokinumab trials and aggregate data from the ADvocate 1 and ADvocate 2 lebrikizumab trials, with IPD weighted to match ADvocate baseline and week-16 characteristics. Week-16 responders, patients achieving 75% reduction in Eczema Area and Severity Index (EASI-75) or Investigator Global Assessment 0 or 1 (IGA 0/1) with an ≥ 2-point improvement from baseline, were re-randomized to active treatment every 2 weeks (Q2W), every 4 weeks (Q4W), or placebo Q2W for the 36-week maintenance period. Week 52 efficacy outcomes included IGA 0/1, EASI-75, 90% reduction in EASI (EASI-90), EASI percentage improvement, and pruritus numerical rating scale 4-point improvement (pruritus 4pt) from baseline. Response differences relative to placebo for endpoints of interest were calculated between tralokinumab and lebrikizumab for Q2W and Q4W dosing to compare maintenance of efficacy at week 52 among week 16-responders.

Results

Differences in efficacy endpoints between tralokinumab and lebrikizumab were not statistically significant; within Q2W dosing, tralokinumab versus lebrikizumab response rate differences at week 52 numerically favored tralokinumab (IGA 0/1: 1.2%, EASI-75: 19.8%, EASI-90: 1.5%, EASI percentage improvement from baseline: 3.3%, pruritus 4pt: 17.6%) while Q4W differences were more variable (IGA 0/1: −20.5%, EASI-75: 15.7%, EASI-90: −8.5%, EASI percentage improvement from baseline: −0.5%, pruritus 4pt: −2.7%).

Conclusions

Comparable maintenance of efficacy was observed between tralokinumab and lebrikizumab after 52 weeks of treatment among patients meeting response criteria at week 16.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13555-025-01563-3.

Key Summary Points

Why carry out this study?

Atopic dermatitis (AD) is a chronic inflammatory skin condition with interleukin-13 as a key driver, and while several biologics targeting this pathway have become available, head-to-head comparisons of their long-term efficacy have yet to be conducted.

This analysis aimed to compare the efficacy of tralokinumab and lebrikizumab at week 52 among patients who met response criteria at week 16 of treatment.

What was learned from this study?

An anchored matching-adjusted indirect comparison (MAIC) analysis of tralokinumab and lebrikizumab revealed no statistically significant differences in efficacy or maintenance of efficacy endpoints at week 52 among week-16 responders, with all endpoints numerically in favor of tralokinumab for biweekly dosing, while treatment every 4 weeks showed no clear numerical preference.

Findings from this anchored MAIC support comparable improvements in AD signs and symptoms between tralokinumab and lebrikizumab at week 52 among week-16 responders, thereby supporting the use of both drugs for long-term management of AD.

Introduction

Atopic dermatitis (AD) is a chronic, relapsing, inflammatory skin condition characterized by redness, pruritus, and eczematous lesions with a considerable impact on patient quality of life [1, 2]. Patients with inadequately controlled disease often experience difficulties in their daily activities, productivity, sleep, and social plans, as well as a higher incidence of anxiety and depression [1–3]. While topical emollients, topical corticosteroids (TCS), and topical calcineurin inhibitors are helpful in mild cases, they are not generally sufficient for moderate-to-severe disease. Patients with moderate-to-severe disease often use systemic oral corticosteroids, the continued use of which is not advised owing to adverse events, including growth impairment, adrenal insufficiency, and increased susceptibility to infection [4–7]. Thus, there is an unmet need for safe, effective long-term therapies for patients with moderate-to-severe disease [8–10]. Biologics have provided a promising avenue for the long-term treatment of AD and are currently recommended as systemic treatment options to prevent flares and maintain disease control for patients with moderate-to-severe disease [3, 5, 6].

The type 2 cytokine interleukin (IL)−13 is a key cytokine involved in skin inflammation, and its expression in the skin is directly correlated to disease severity in AD [11–13]. Tralokinumab and lebrikizumab are two monoclonal antibodies that have been developed to target the IL-13 signaling pathway in AD; these monoclonal antibodies bind different epitopes and inhibit distinct steps in the signal transduction pathway (i.e., preventing free IL-13 from binding IL-13Rα1/2 or by preventing the IL-13/IL-13Rα1 complex from heterodimerizing with IL-4Rα, respectively) [11, 13, 14]. Several monoclonal antibodies have been developed to target other type 2 inflammatory pathways, including dupilumab (targeting both IL-4 and IL-13 signaling) and nemolizumab (targeting IL-31 signaling) [11, 15]. Phase 3 studies for these biologics in patients with moderate-to-severe AD have demonstrated their efficacy at week 16, including reduction in Investigator Global Assessment score to 0 or 1 (IGA 0/1), indicating clear or almost clear skin, in 15.8–43.1% of patients and 75% reduction in Eczema Area and Severity Index (EASI-75), achieved by 25.0–58.8% of patients and with continued benefit through week 52 [15–21]. However, direct comparisons of these drugs have yet to be conducted. Without head-to-head data, indirect comparison methods such as Bucher, network meta-analysis (NMA), simulated treatment comparisons, or matching-adjusted indirect comparison (MAIC) can be used. An anchored MAIC can be the most appropriate indirect comparison method in some cases, as it allows for individual patient data (IPD) from one trial to be compared with summary data of another by matching summary statistics and relying on a common comparator arm (e.g., placebo) to minimize the influence of potential effect modifiers [22]. A previous MAIC analysis comparing the efficacy of tralokinumab and dupilumab at week 32 demonstrated similar efficacy across endpoints [23]; however, there have been no indirect comparisons of long-term treatment with tralokinumab and lebrikizumab.

To compare the maintenance of efficacy of tralokinumab and lebrikizumab at week 52, an anchored MAIC was performed using the ECZTRA 1 and ECZTRA 2 (ECZTRA 1 & 2) tralokinumab and ADvocate 1 and ADvocate 2 (ADvocate 1 & 2) lebrikizumab monotherapy trials owing to their similar designs and longer durations compared with available TCS combination trials. As AD is a chronic, lifelong disease, and early response rates at week 16 may not be indicative of long-term efficacy [24], comparisons were made at week 52 among patients who met response criteria on active treatment at week 16 and were re-randomized to active treatment every 2 weeks (Q2W), every 4 weeks (Q4W), or placebo Q2W for the maintenance period for each respective trial.

Methods

MAIC Source Data

This anchored MAIC compared the maintenance of efficacy of tralokinumab and lebrikizumab using IPD from the ECZTRA 1 & 2 (NCT03131648 and NCT03160885) tralokinumab trials and aggregate data from the ADvocate 1 & 2 (NCT04146363 and NCT04178967) lebrikizumab trials [16–19, 22, 25, 26]. A summary of these trials and their key differences are shown in Fig. 1 and presented in Table 1, respectively.

Fig. 1.

ECZTRA 1 & 2 and ADvocate 1 & 2 trial design. Patients were deemed week-16 responders if IGA 0/1 with ≥ 2-point improvement or EASI-75 were achieved. Week-16 responders were re-randomized 2:2:1 to active treatment Q2W, Q4W, or placebo Q2W for an additional 36-week maintenance period. Patients were transferred to the open-label arm if there was a ≥ 2-point increase in IGA or lack of EASI-75 response for 4 weeks in ECZTRA 1 & 2 or lack of EASI-50 maintenance in ADvocate 1 & 2. EASI Eczema Area and Severity Index, EASI-50 50% reduction in EASI, EASI-75 75% reduction in EASI, EASI-90 90% reduction in EASI, IGA Investigator Global Assessment, LD loading dose, MAIC matching-adjusted indirect comparison, Q2W every 2 weeks, Q4W every 4 weeks, R randomized, TCS topical corticosteroids

Table 1.

Key study characteristics of ECZTRA 1 & 2 and ADvocate 1 & 2

	ECZTRA 1 & 2 Tralokinumab 300 mg	ADvocate 1 & 2 Lebrikizumab 250 mg
Design	Randomized, double‐blind, multicenter, placebo‐controlled phase 3	Randomized, double-blind, multicenter, placebo-controlled phase 3
Loading dose	600 mg on day 0	500 mg at day 0 and week 2
Inclusion criteria	• Age ≥ 18 years with moderate-to-severe AD • History of AD ≥ 1 year • Inadequate response to topical therapy or topicals not advised • EASI ≥ 12 at screening and ≥ 16 at baseline • IGA ≥ 3 • AD involvement of ≥ 10% of the body surface area • Worst daily pruritus NRS ≥ 4 average	• Adults (≥ 18 years) and adolescents (12–18 years, ≥ 40 kg) with moderate-to-severe AD • History of AD ≥ 1 year • Inadequate response to topical therapy or topicals not advised • EASI ≥ 16 • IGA ≥ 3 • AD involvement of ≥ 10% of the body surface area
Exclusion criteria	• Previous enrollment in a tralokinumab trial • Active conditions that may confound AD diagnosis • Active conditions that may confound AD diagnosis • Treatment within the previous 4 weeks of a systemic immunosuppressant/immunomodulator, systemic corticosteroid, or 3+ bleach baths in any week • Treatment with TCS, TCI, or topical PDE-4 inhibitor for 2 weeks prior to randomization (washout period)	• Previous treatment with lebrikizumab, dupilumab, or tralokinumab • Treatment within the previous 4 weeks of an immunosuppressant/immunomodulator or photochemotherapy • Use of an investigational drug within 5 half-lives of baseline • Uncontrolled chronic disease that may require oral corticosteroids • Treatment with TCS, TCI, or topical PDE-4 inhibitor for 1 week prior to baseline visit (washout period)
TCS during maintenance period	Used at the discretion of the investigator for intolerable symptoms	Intermittent use of TCS permitted
Transfer to open-label arm	≥ 2-point increase in IGA and/or lack of EASI-75 response for 4 weeks (three consecutive visits)a	Lack of EASI-50 response
Patient data included in the present MAIC analysis	Week 52 IPD of patients initially randomized to tralokinumab who met response criteria at week 16 (IGA 0/1 and/or EASI-75)	Week 52 aggregate data containing patients initially randomized to lebrikizumab who met response criteria at week 16 (IGA 0/1 and/or EASI-75)

^aSpecifically, patients with IGA 0 at week 16 who had IGA ≥ 2 and did not achieve EASI-75 over at least a 4-week period (i.e., over three consecutive visits), patients with IGA 1 at week 16 who had IGA ≥ 3 and did not achieve EASI-75 over at least a 4-week period, or patients with IGA > 1 at week 16 who did not achieve EASI-75 over at least a 4-week period were transferred to open-label treatment (tralokinumab 300 mg Q2W with optional use of TCS) and continued their scheduled visit sequence. Transfer to the open-label arm occurred no earlier than week 22

Statements in bold indicate key differences between the included trials

AD atopic dermatitis, EASI Eczema Area and Severity Index, EASI-50 50% reduction in EASI, EASI-75 75% reduction in EASI, IGA Investigator Global Assessment, IPD individual patient data, LOCF last observation carried forward, NRS numerical rating scale, PDE phosphodiesterase, TCI topical calcineurin inhibitors, TCS topical corticosteroids

ECZTRA 1 & 2 were randomized, double-blind, multicenter, placebo-controlled phase 3 trials with patients randomized 3:1 to receive tralokinumab 300 mg Q2W after a 600-mg loading dose on day 0, or placebo Q2W during the 16-week initial treatment phase [16, 18] (Fig. 1, Table 1). Responders at week 16, defined as patients who achieved IGA 0/1 and/or EASI-75, were re-randomized 2:2:1 to receive tralokinumab 300 mg Q2W, tralokinumab 300 mg Q4W, or placebo Q2W for an additional 36-week maintenance period. Responders in the placebo arm continued to receive placebo Q2W during the maintenance period. Nonresponders were transferred to the open-label arm, where they received tralokinumab 300 mg Q2W with optional TCS until week 52. Week-16 responders who failed to maintain response during the maintenance period were transferred to the open-label arm. Failure to maintain response was defined as either a ≥ 2-point worsening in IGA score or failing to maintain an EASI-75 response for three or more consecutive visits. Rescue treatment, including TCS, during the maintenance period was permitted under the discretion of the investigator.

ADvocate 1 & 2 were randomized, double-blind, multicenter, placebo-controlled phase 3 trials with patients randomized 2:1 to receive lebrikizumab 250 mg Q2W after a 500-mg loading dose at day 0 and week 2, or placebo Q2W during the 16-week initial treatment phase [17, 19] (Fig. 1, Table 1). Patients who achieved clinical response (IGA 0/1 and/or EASI-75) at week 16 on either active treatment or placebo were re-randomized 2:2:1 to receive lebrikizumab 250 mg Q2W, lebrikizumab 250 mg Q4W, or placebo Q2W for an additional 36-week maintenance period, with intermittent TCS use permitted. Patients who did not achieve clinical response at week 16 and patients who did not maintain a 50% reduction in EASI (EASI-50) during the maintenance period were transferred to the open-label arm and received lebrikizumab 250 mg Q2W with intermittent TCS use permitted.

These MAIC analyses utilized week-52 data from patients with available baseline data who were initially randomized to active treatment, met response criteria at week 16 without using rescue medication, and were re-randomized to maintenance treatment in their respective trials (Fig. 1). IPD for week-16 responders in the ECZTRA 1 & 2 trials were selected by applying the inclusion criteria from ADvocate 1 & 2 and were weighted to match the baseline and week-16 characteristics of ADvocate 1 & 2 patients. The baseline characteristics matched were mean age, sex, mean body mass index (BMI), mean AD duration, race, mean EASI, mean Dermatology Life Quality Index (DLQI), mean worst daily pruritus numerical rating scale (NRS), and proportion of patients with IGA 3. The week-16 characteristics matched were mean EASI, proportion of patients with IGA 0/1, and mean pruritus NRS. Only patients who had available data for all baseline and week-16 matching parameters were included in the analysis.

Outcome Measures

Maintenance of efficacy at week 52 was measured by EASI-75, pruritus NRS ≥ 4-point improvement, and IGA 0/1 with ≥ 2-point improvement specifically among patients who had met each respective measure at week 16. Efficacy at week 52 was measured by 90% reduction in EASI (EASI-90) and EASI percentage improvement from baseline, both of which were assessed in patients with EASI-75 response at week 16. All measures of improvement were presented relative to baseline.

Statistical Analysis

The analyses were performed using a nonresponder estimand. Patients who received rescue medication such as TCS or topical calcineurin inhibitors, discontinued treatment, or were transferred to the open-label arm were imputed as nonresponse for binary endpoints and imputed as last observation carried forward (LOCF) for continuous endpoints. All statistical tests were performed using SAS, version 9.4 (SAS Institute Inc., Cary, USA), and all tests were two-sided with an alpha level of 0.05.

The ADvocate 1 & 2 trials reported two estimands: the nonresponder estimand and a hybrid estimand. For the hybrid estimand, patients who received systemic rescue medication, discontinued owing to lack of efficacy, or were transferred to the open-label arm had baseline values carried forward after this time, and patients who received topical rescue medication or discontinued treatment for any other reason had the missing values imputed using Markov Chain Monte Carlo imputation. A sensitivity analysis was performed by using the hybrid estimand and comparing the results with results based on the nonresponder estimand.

Ethical Approval

This analysis is based on previously conducted studies and does not contain data from new studies with human participants; therefore, ethical approval was not required.

Results

Matching of Baseline and Week-16 Characteristics

In ECZTRA 1 & 2, 1192 patients received tralokinumab 300 mg Q2W during the 16-week initial treatment period [16, 18]. Of patients who met trial response criteria at week 16 and were re-randomized, 328 met inclusion criteria for this analysis, with 127 patients in the tralokinumab Q2W arm, 130 patients in the tralokinumab Q4W arm, and 71 patients in the placebo Q2W arm for the maintenance period. After matching and weighting of the IPD, the effective sample sizes from ECZTRA 1 & 2 were 71 in the tralokinumab Q2W arm, 68 in the tralokinumab Q4W arm, and 28 in the placebo Q2W arm (Table 2). In ADvocate 1 & 2, 564 patients were initially randomized to receive lebrikizumab 250 mg Q2W, with 291 responders at week 16 [17, 27]. These responders were re-randomized with 113 patients in the lebrikizumab Q2W arm, 118 patients in the lebrikizumab Q4W arm, and 60 patients in the placebo arm [27].

Table 2.

Baseline and week-16 characteristics in ADvocate 1 & 2 and ECZTRA 1 & 2 trials and weighted groups

		ADvocate 1 & 2 [19]	ECZTRA 1 & 2		ADvocate 1 & 2 [19]	ECZTRA 1 & 2		ADvocate 1 & 2 [19]	ECZTRA 1 & 2
		Placebo	Placebo		Lebrikizumab Q2W	Tralokinumab Q2W		Lebrikizumab Q4W	Tralokinumab Q4W
		Baseline characteristics	Baseline characteristics	Weighted summary	Baseline characteristics	Baseline characteristics	Weighted summary	Baseline characteristics	Baseline characteristics	Weighted summary
		N = 60	N = 71	Neff  = 28	N = 113	N = 127	Neff  = 71	N = 118	N = 130	Neff  = 68
Baseline	Age, years (SD)	34 (17)	38 (13)	34 (12)	36 (17)	37 (13)	36 (13)	36 (17)	39 (16)	36 (15)
	Sex, % male	40	56	40	53	57	53	42	51	42
	Mean BMI, kg/m2 (SD)	25 (5)	26 (6)	25 (6)	26 (7)	26 (5)	26 (6)	26 (6)	26 (6)	26 (6)
	Disease duration, years (SD)	20 (15)	30 (16)	20 (13)	22 (14)	27 (16)	22 (14)	23 (15)	27 (16)	23 (15)
	Race, % white	55	62	55	71	68	71	73	70	73
	IGA 3, %	62	62	62	62	72	62	66	56	66
	Mean EASI (SD)	29 (11)	29 (12)	29 (15)	30 (11)	28 (12)	30 (13)	29 (13)	28 (12)	29 (13)
	Mean DLQI (SD)	15 (8)	17 (6)	15 (6)	15 (7)	16 (8)	15 (7)	15 (8)	15 (7)	15 (7)
	Mean pruritus NRS (SD)	8 (2)	8 (1)	8 (1)	7 (2)	8 (2)	7 (1)	7 (2)	8 (1)	7 (2)
Week 16	Mean EASI (SD)	2 (2)	3 (3)	2 (2)	3 (3)	3 (3)	3 (3)	2 (3)	3 (3)	2 (2)
	IGA 0/1, %	67	65	67	68	72	68	67	62	67
	Mean pruritus NRS (SD)	3 (2)	4 (3)	3 (2)	3 (2)	3 (2)	3 (2)	3 (2)	4 (3)	3 (2)

DLQI Dermatology Life Quality Index, EASI Eczema Area and Severity Index, IGA Investigator Global Assessment, N sample size, N_eff effective sample size after adjusted matching, NRS numerical rating scale, Q2W every 2 weeks, Q4W every 4 weeks, SD standard deviation

Though baseline characteristics were generally well balanced between trials, some modest differences were noted prior to matching (Table 2). Across all maintenance treatment groups, baseline disease duration was longer in ECZTRA 1 & 2 when compared with ADvocate 1 & 2. The proportion of patients with IGA 3 (moderate disease) at baseline was also higher in ECZTRA 1 & 2 for the Q2W arm but lower in the Q4W arm compared with ADvocate 1 & 2. The mean EASI score at baseline was comparable across the two trials for both dosing schedules and the placebo groups.

Tralokinumab and Lebrikizumab Had Comparable Maintenance of Efficacy Across Endpoints at Week 52

Among week-16 responders re-randomized to Q2W dosing, there were no statistically significant differences for the five efficacy endpoints when comparing tralokinumab and lebrikizumab at week 52; however, all matched point estimates were numerically in favor of tralokinumab (Fig. 2; unmatched results shown in Supplementary Fig. S1). The matched response rate differences measuring clinician-assessed endpoints between tralokinumab and lebrikizumab were IGA 0/1 with ≥ 2-point improvement of 1.2% (95% confidence interval [CI] −31.0% to 33.5%; P > 0.05), EASI-75 of 19.8% (95% CI −5.9% to 45.4%; P > 0.05), EASI-90 of 1.5% (95% CI −24.0% to 27.1%; P > 0.05), and EASI percentage improvement from baseline of 3.3% (95% CI −5.5% to 12.0%; P > 0.05). The matched response rate difference for the patient-reported outcome of pruritus NRS ≥ 4-point improvement was 17.6% (95% CI −18.4% to 53.5%; P > 0.05).

Fig. 2.

Matched response difference from placebo in achieving efficacy endpoints for tralokinumab versus lebrikizumab at week 52 among week-16 responders re-randomized to Q2W dosing. *A P-value of ≤ 0.05 was considered statistically significant. P-value is for two-sided test of zero difference between responder proportions for tralokinumab versus lebrikizumab. ECZTRA treatment differences (tralokinumab versus placebo) are unadjusted (and weighted) risk differences with asymptotic standard errors. The applied weights are derived so that baseline means for selected variables match comparator data. ADvocate treatment differences (lebrikizumab versus placebo) are derived from published data [19]. The indirect comparisons are calculated as ECZTRA treatment differences minus ADvocate treatment differences and are thus anchored in placebo. Standard errors for the indirect comparisons are calculated as the square root of the sum of respective treatment difference variances. Within-treatment ECZTRA proportions are estimated by back-transformation from logit models. Prior to analysis, ECZTRA data following rescue medication, discontinuation, or transfer to open-label arm as well as missing data have been imputed as nonresponse. CI confidence interval, EASI Eczema Area and Severity Index, EASI-75 75% reduction in EASI, EASI-90 90% reduction in EASI, IGA Investigator Global Assessment, lebri lebrikizumab, pt point, MAIC matching-adjusted indirect comparison, n number of patients included, NRS numerical rating scale, Q2W every 2 weeks, RD response difference, RR response rate, SE standard error, tralo tralokinumab

There were also no statistically significant differences for responders re-randomized to Q4W dosing at week 16, with numerical preferences toward either tralokinumab or lebrikizumab varying by parameter for each matched point estimate (Fig. 3; unmatched results shown in Supplementary Fig. S2). When comparing matched response rate differences between tralokinumab and lebrikizumab, disease severity as measured by IGA 0/1 with ≥ 2-point improvement was −20.5% (95% CI −52.8% to 11.9%; P > 0.05) and EASI-90 was −8.5% (95% CI −34.0% to 16.9%; P > 0.05), with both numerically in favor of lebrikizumab. The matched response rate difference for EASI-75 was 15.7% (95% CI −9.9% to 41.3%; P > 0.05), numerically in favor of tralokinumab. EASI percentage improvement from baseline was −0.5% (95% CI −9.2% to 8.2%; P > 0.05), and pruritus NRS ≥ 4-point improvement was −2.7% (95% CI −38.3% to 32.9%; P > 0.05), suggesting comparable maintenance of efficacy between the two drugs.

Fig. 3.

Matched response difference from placebo in achieving efficacy endpoints for tralokinumab versus lebrikizumab at week 52 among week-16 responders re-randomized to Q4W dosing. *A P-value of ≤ 0.05 was considered statistically significant. P-value is for two-sided test of zero difference between responder proportions for tralokinumab versus lebrikizumab. ECZTRA treatment differences (tralokinumab versus placebo) are unadjusted (and weighted) risk differences with asymptotic standard errors. The applied weights are derived so that baseline means for selected variables match comparator data. ADvocate treatment differences (lebrikizumab versus placebo) are derived from published data [19]. The indirect comparisons are calculated as ECZTRA treatment differences minus ADvocate treatment differences and are thus anchored in placebo. Standard errors for the indirect comparisons are calculated as the square root of the sum of respective treatment difference variances. Within-treatment ECZTRA proportions are estimated by back-transformation from logit models. Prior to analysis, ECZTRA data following rescue medication, discontinuation, or transfer to open-label arm as well as missing data have been imputed as nonresponse. CI confidence interval, EASI Eczema Area and Severity Index, EASI-75 75% reduction in EASI, EASI-90 90% reduction in EASI, IGA Investigator Global Assessment, lebri lebrikizumab, pt point, MAIC matching-adjusted indirect comparison, n number of patients included, NRS numerical rating scale, Q4W every 4 weeks, RD response difference, RR response rate, SE standard error, tralo tralokinumab

In support of these results, the sensitivity analysis showed no statistically significant differences between tralokinumab and lebrikizumab in either dosing schedule (Supplementary Fig. S3).

Discussion

Currently available therapies for adults with moderate-to-severe AD with inadequate responses to topical or systemic therapies include Janus kinase (JAK) inhibitors and biologics such as dupilumab, nemolizumab, tralokinumab, and lebrikizumab [3, 5, 11, 28], but head-to-head comparisons of these options have been limited. Given the chronic nature of AD, it is imperative that therapeutic strategies provide enduring relief and long-term disease management. We have, therefore, conducted an anchored MAIC with the objective of comparing the efficacy and maintenance of efficacy of tralokinumab and lebrikizumab at week 52 in patients meeting the predefined response criteria at week 16. The comparisons presented here highlight the sustained long-term efficacy of both tralokinumab and lebrikizumab, a crucial quality of effective treatments for AD. Our analyses indicate that tralokinumab and lebrikizumab have comparable maintenance of efficacy across all endpoints at week 52 in patients who initially responded to active treatment at week 16. Although no differences relative to placebo were statistically significant, response differences were numerically in favor of tralokinumab Q2W for all evaluated endpoints, while individual Q4W endpoints were numerically in favor of either drug or suggested equivalence.

There were several limitations to our study. One limitation includes the numerous differences in trial design between the ECZTRA and ADvocate trials (Table 1). Differences between phase 3 trials, such as loading-dose frequency, inclusion of adolescents, duration of washout period, TCS use, and criteria to transfer to the open-label arm, all hold the potential to influence study results and subsequently make it difficult to compare results [24, 29]. For this analysis, complete population matching was not possible owing to differences in included age groups between trials. While biologics have generally shown efficacy in pediatric and adult populations with AD, there may be differences in disease phenotypes, treatment response rates, and safety profiles [30–32]. In addition, the duration of the washout period can impact downstream elements of the trial through residual carryover treatment effects, possibly impacting measured disease severity at baseline and rescue treatment use [29]. The differences in criteria dictating transfer to open-label treatment can also influence who is included in the final data set. In the ADvocate trials, patients were transferred to the open-label arm if they did not maintain an EASI-50 response, whereas in the ECZTRA trials, patients were transferred to open-label treatment if they did not maintain an EASI-75 response for three consecutive visits or if their IGA score worsened by 2 or more points from week 16. Therefore, ADvocate patients were able to stay in the maintenance arms longer before being required to transfer to the open-label arm, potentially allowing time to regain response. As with all indirect comparisons, bias due to observed and unobserved differences across the trials cannot be ruled out.

As patients maintain response to treatment over time, patients and their providers may be interested in adjusting dosing schedules with the goal of reducing the number of injections and subsequently improving convenience for patients. A recent report demonstrated that 72.2% of patients achieving stable disease control on tralokinumab Q2W (IGA 0/1 and worst daily pruritus NRS < 3 over weeks 12–16) were able to maintain response when switched to less frequent Q4W dosing, and that over 90% of patients who relapsed after switching to Q4W were able to regain response by reverting to Q2W dosing [33]. Both Q2W and Q4W are approved maintenance dosing schedules of tralokinumab, with patients being able to switch to Q4W once achieving disease control [34, 35]. According to the European Medicines Agency (EMA) label for lebrikizumab, after the initial loading doses, lebrikizumab 250 mg is given Q2W until week 16, and by week 24, dosing is updated strictly to Q4W, while the US Food and Drug Administration (FDA) label permits this change to occur upon later adequate clinical response [36, 37]. Therefore, tralokinumab provides a unique dosing flexibility for a more personalized treatment approach to the long-term treatment of flare-driven disease.

In the absence of head-to-head data, there are several options to compare results between trials, and for this analysis, an anchored MAIC was the most appropriate. Bucher and NMA indirect comparisons do not match patient populations and therefore can introduce bias [38], and unanchored simulated treatment comparisons or unanchored MAICs can introduce bias owing to differences in effect modifiers or prognostic factors [39]. An anchored MAIC adjusts for differences in study design, and is suitable to compare the efficacy of therapeutics using IPD from one trial and aggregate patient data from another trial in a specific patient population and relies on a common comparator arm (e.g., placebo) [22, 26]. This method requires matching of the summary statistics of demographic and disease characteristics deemed clinically relevant for parameters available from both trials, mitigating the influence of potential effect modifiers that may have influenced treatment-related outcomes. In an anchored MAIC, the use of a common comparator arm reduces bias from unmatched prognostic factors that may influence all arms equally, thus providing a more robust and reliable comparison than an unanchored MAIC that does not require a common comparator arm.

A previously conducted unanchored MAIC comparing tralokinumab and dupilumab, both plus optional TCS, at week 32 of treatment showed a similar efficacy across IGA 0/1, EASI-50, EASI-75, and EASI-90 responses, and a larger improvement in DLQI from baseline with tralokinumab, which was statistically significant [23]. The anchored MAIC conducted here adds to these findings by supporting comparable maintenance of efficacy between tralokinumab and lebrikizumab at week 52 among week-16 responders, with both offering similar long-term benefits for patients with AD, therefore supporting their use for managing the chronic nature of the disease. These findings underscore the importance of considering available biologics as effective long-term treatment strategies for AD, with the choice of biologics potentially guided by individual patient preferences and other differences such as dose flexibility, safety profile, or observations from clinical practice rather than significant differences in efficacy. Further direct comparative studies and real-world analyses from patient registries are warranted to confirm these results and to explore any potential differences in safety profiles and patient-reported outcomes.

Conclusions

The anchored MAIC analysis conducted between the ECZTRA 1 & 2 trials for tralokinumab and ADvocate 1 & 2 trials for lebrikizumab provides valuable insights into the comparison of long-term efficacy of these biologics in treating AD. The analyses revealed that both tralokinumab and lebrikizumab demonstrated comparable maintenance of efficacy at week 52 among patients who met clinical response criteria at week 16. While the biweekly dosing regimen of tralokinumab showed a numerical advantage over lebrikizumab across most analysis endpoints, these differences did not reach statistical significance.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Matthias Augustin, April Armstrong, Naiem T. Issa, Anne Sohrt Petersen, Rie von Eyben, Teodora Festini, and Tiago Torres all contributed to analysis conception and design. Data analyses were performed by Rie von Eyben. All authors reviewed and revised manuscript drafts. All authors read and approved the final manuscript.

Funding

Sponsorship for this analysis and the Rapid Service Fee for Dermatology and Therapy were funded by LEO Pharma A/S (Ballerup, Denmark).

Data Availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.

Declarations

Conflict of Interest

Matthias Augustin has served as a consultant, lecturer, researcher, and/or has received institutional research grants from companies manufacturing drugs for atopic dermatitis, including AbbVie, Almirall, Bayer, Beiersdorf, Eli Lilly, Galderma, Incyte, LEO Pharma, L’Oreal, MSD, Novartis, Pfizer, Regeneron, Roche-Posay, Sanofi-Genzyme, and Sun. April Armstrong has served as a consultant for and received honoraria from AbbVie, Almirall, Arcutis, ASLAN, Beiersdorf, Boehringer Ingelheim, Bristol Myers Squibb, Dermavant, Dermira, EPI, Incyte, Janssen, LEO Pharma, Eli Lilly, Nimbus, Novartis, Ortho Dermatologics, Pfizer, Regeneron, Sanofi, Sun, and UCB and has participated in advisory boards for Boehringer Ingelheim and Parexel. Naiem T. Issa has received funding from the following entities either as a speaker, consultant, advisor, or investigator: Abbvie, Almirall, Apogee, Bristol Myers Squibb, Castle Biosciences, Dermavant Sciences, DermTech, Galderma, Incyte, Janssen, Journey, LEO Pharma, Lilly, Ortho Dermatologics, Pfizer, Primus, Regeneron, Sanofi, SUN Pharmaceuticals Industry, Topix, UCB, and Verrica Pharmaceuticals. Anne Sohrt Petersen was an employee of LEO Pharma. Rie von Eyben is an employee of LEO Pharma. Teodora Festini is an employee of LEO Pharma. Tiago Torres has received consultancy and/or speaker’s honoraria from and/or participated in clinical trials sponsored by AbbVie, Almirall, Amgen, Arena Pharmaceuticals, Biocad, Biogen, Boehringer Ingelheim, Bristol Myers Squibb, Celgene, Fresenius Kabi, Janssen, LEO Pharma, Eli Lilly, MSD, Mylan, Novartis, Pfizer, Samsung-Bioepis, Sandoz, Sanofi, and UCB.

Ethical Approval

This analysis is based on previously conducted studies and does not contain data from new studies with human participants; therefore, ethical approval was not required.