AD-REAL: A Prospective, Multinational, Observational Cohort Study of Clinical Effectiveness and Treatment Discontinuation of Oral Systemics for Moderate-to-Severe Atopic Dermatitis Authors: Matthias Augustin, Andreas Pinter, Silvia Sabatino, Louis P Sanden, Pablo Chicharro, Esther Serra-Baldrich, Mahreen Ameen, Anthony Bewley, Walid Fakhouri, Inmaculada de la Torre, Julien Seneschal Published: 2026 ABSTRACT AbstractIntroductionClinical data surrounding treatment patterns and reasons for discontinuation of oral systemics (OS) for the treatment of adult patients with moderate–severe atopic dermatitis (AD) remain limited. AD-REAL descriptively reports the discontinuation rates and treatment responses at week (W) 24 for patients with AD initiating new OS treatments.MethodsAD-REAL is a 12-month, multinational, observational cohort study of adult patients with moderate-to-severe AD commencing either conventional systemics (CS) or baricitinib. Other Janus kinase inhibitors (JAKis) were analyzed ad hoc. The primary objective reported the discontinuation rate of patients at W24 from baseline. Secondary objectives reported outcomes for Eczema Area and Severity index (EASI), Itch Numerical Rating Scale (NRS), and pain NRS.ResultsAt baseline (N = 313), patients initiated CS (N = 113), baricitinib (N = 87), and other JAKi (N = 92), with a mean EASI of 19.8, 15.8, and 16.4, respectively. At W24, the Kaplan–Meier cumulative incidence of initial treatment discontinuation was 50.8% (95% confidence interval [95% CI] 40.9, 59.8) for CS, 29.0% (95% CI 19.8, 38.8) for baricitinib, and 18.8% (95% CI 11.5, 27.5) for other JAKi. The primary reason for discontinuation of baricitinib and other JAKi was primary lack of effectiveness, while CS discontinued owing to the addition of concomitant systemic treatment, adverse events, and patient/physician decisions. Of the baricitinib and other JAKi cohorts, 81.0% and 84.2% achieved an EASI ≤ 7 score, 38.1% and 38.2% improvement in pain, and 42.0% and 38.5% improvement in itch, while 63.1%, 16.7%, and 14.3% achieved these outcomes for CS, respectively.ConclusionsIn this descriptive observational cohort study, patients treated with baricitinib and other JAKi reported numerically lower discontinuation rates and numerically improved clinical outcomes than patients treated with CS. Similar to early reports of baricitinib effectiveness in treating an itch-dominant subpopulation of patients with AD, patients receiving baricitinib in clinical practice reported improved outcomes for severe itch.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s13555-026-01852-5.Keywords: Atopic dermatitis, Oral systemics, Observational study, Baricitinib Key Summary Points Why carry out this study? Oral systemics (OS) are used to treat atopic dermatitis (AD). However, in AD, limited information is available from randomized clinical trials (RCTs) and real-world studies regarding conventional OS therapies, their duration, and the reasons for patient discontinuation.Despite new OS entering the market with RCT and on-label indication for first-line systemic therapy, conventional systemics (CS) remain broadly used for AD in many countries.This represents a major knowledge gap in the management of adult patients with AD. Therefore, AD-REAL aimed to address this gap by descriptively reporting on the use of CS treatments, baricitinib, and other JAKis approved for AD across four European countries. What was learned from the study? After 24 weeks of treatment, patients treated with baricitinib and other JAKi reported numerically lower discontinued rates and numerically improved clinical outcomes than the CS cohort.Patients treated with CS discontinued treatment primarily due to adverse events or the addition of a concomitant immunomodulatory drug, while patients treated with baricitinib or other JAKi discontinued treatment due to a lack of treatment effectiveness.Patients receiving baricitinib reported improvement in severe itch, similar to previous reports of its effectiveness in an itch-dominant subpopulation of patients with AD. Introduction AD is a common chronic inflammatory skin disorder, with heterogeneous clinical symptoms and signs including itch, and lichenification [1, 2]. Within the adult population, the prevalence of AD varies from 2% to 8% worldwide, with one-third of cases occurring in adulthood [3, 4]. In Europe, the prevalence among adults ranges between 2.2% and 17.6%, depending on the country. Across regions, AD poses a significant burden on patient quality of life (QOL); this burden is exacerbated in patients whose AD symptoms are inadequately controlled [5]. AD is also known to present with frequent comorbidities, such as sleep disturbances and depression, regardless of the severity of AD symptoms, which further negatively impact patient QOL [6]. As a result of the high disease burden, patients show a large number and variety of therapeutic goals [7]. Likely owing to disease heterogeneity, therapeutic responses are known to vary between patients with AD [4, 8, 9]. In Europe, the standard of care for patients with AD focuses on the management of AD flares by topical treatment, phototherapy, and systemic therapies, with avoidance strategies being used to limit trigger exposures [4, 10]. Prior to the recent licensed approval of biologic (dupilumab, lebrikizumab, and tralokinumab) and oral JAKi (baricitinib, abrocitinib, and upadacitinib) systemics, cyclosporine was the only conventional OS approved in the EU for the treatment of AD, with the exception of steroids [10, 11]. However, this approval was limited to severe AD, while therapeutics such as baricitinib were approved for adult patients with moderate-to-severe AD [10, 11]. Although cyclosporine can be an effective treatment, its long-term use is limited by its side effects that require blood pressure and renal function follow-ups. Consensus-based European guidelines recommend that treatment with cyclosporine should not exceed a continuous regimen of 2 years, while 3–6-month intervals of dosages between 2.5 mg/kg/day and 3.0 mg/kg/day are recommended during maintenance [3, 4, 10]. Other off-label conventional OS, including methotrexate, mycophenolate mofetil, azathioprine, and systemic corticosteroids, are prescribed on the basis of clinical guidelines, but their long-term use is also limited by poor tolerability [10–12]. Moreover, data from RCTs and real-world studies of these therapies, their treatment duration, and reasons for discontinuation in adult patients with AD are limited. Therefore, evidence from clinical practice describing the management of AD with OS therapies represents a major knowledge gap. However, real-world evidence for the current clinical practices and treatment success of OS in patients with AD is limited. The AD-REAL observational study [13] is the first to report the use of oral CS treatments in real-world clinical practice across four European countries (France, Germany, Spain, and the UK). As previously reported, the majority of patients who participated in AD-REAL had long-standing AD, and nearly one-third were refractory to systemic therapies at baseline, reporting previous treatment failure [13]. More than one-third of these patients presented with a body surface area (BSA) ≤ 40% and itch numerical rating scale (NRS) ≥ 7 score, with > 60% prescribed JAKi treatment. This study was designed to descriptively outline the real-world discontinuation pattern and clinical effectiveness of CS and baricitinib in patients who commenced new treatment regimens. Herein, we report the primary and secondary outcomes at Week 24 (W24) from baseline. Data for other OS, including other JAKi, approved during the study enrollment and per protocol inclusion criteria are also reported. Methods Study Design AD-REAL is a 12-month prospective, multicenter, international, observational, non-interventional, and non-randomized cohort study approved on 23 October 2020. It included patients from four European countries (France, Germany, Spain, and the UK). Owing to the COVID-19 pandemic, enrollment was delayed and extended between 16 December 2021 and 21 December 2023. Data were collected at baseline and included post-baseline follow-up visits at weeks 4 (± 1 week), 12 (± 2 weeks), 24 (± 7 weeks), and at month 12 (± 7 weeks). This observational study was conducted in accordance with the ethical principles that have their origin in the Declaration of Helsinki and that are consistent with Good Pharmacoepidemiology Practices and applicable laws and regulations of the country or countries where the study is being conducted, as appropriate. The study protocol was approved by the Institutional Review Board or Ethics Committee at each participating site, which included Le comité de protection des personnes Île-de-France II (N/A), Ethic Kommission der Arztekammer Hamburg (#2021-10562-BO-ff), Hospital Clínico San Carlos (#22/124-O_M_SP), and North of Scotland Research Ethics Committee (#21/NS/0085), and HRA and Health and Care Research Wales (#21/NS/0085). Written informed consent was obtained from all eligible participants before undergoing study-related procedures. Study Patients Data were evaluated from male and female patients (≥ 18 years) with AD at least 6 months prior to baseline and who were eligible for administration of oral CS and other OS. Patients were either initiating OS for the first time (systemic naïve) or had received OS previously and were initiating a new OS treatment (systemic experience). Concomitant systemic therapies were not permitted at baseline but were allowed after study initiation. Concomitant topical therapies were permitted at baseline. The overall treatment cohorts evaluated were baricitinib and CS prescribed at baseline as part of the standard care regimen, including, but not limited to, cyclosporin, methotrexate, mycophenolate mofetil, and azathioprine. Eligible treatments are those that became available per country throughout the enrollment period. Any additional OS treatment that became available during the study period and was reimbursed by country-specific authorities were included in the study but not in the post-baseline analysis. In the case of other JAKis (abrocitinib and upadacitinib), which became available during the delayed enrollment period, this group was included as an exploratory post-baseline analysis in this study. A subgroup analysis was also conducted on patients stratified by BSA and itch NRS. Exclusion criteria included completion or previous discontinuation from the study, active participation in any other study with investigational products, changes to the dosing or administration intervals of other OS, and the re-initiation of another OS treatment in the 4 weeks preceding the commencement of the study. Primary and Secondary Objectives at 24 Weeks The primary objective of this observational study is to report descriptively the proportion of patients with all-cause treatment discontinuation at W24 after initiation of a new OS treatment at baseline for the baricitinib and the CS treatment cohorts. Treatment discontinuation was defined as the discontinuation of the OS treatment initiated at baseline, switching from other OS treatment initiated at baseline to another OS or biologic treatment, treatment interruption ≥ 30 days (initiation, stopping, and re-initiation of the same treatment), and the addition of a concomitant systemic immunomodulatory treatment to control skin symptoms. Patients who received a concomitant systemic immunomodulatory treatment after baseline were considered non-responders from the time that the concomitant treatment started and were censored. Specific reasons for discontinuation included primary lack of effectiveness, secondary loss of effectiveness, adverse events, lack of compliance, patient/physician decision, AD remission, and “other.” Data were collected in the AD treatment form of the electronic case report forms (eCRFs), with start and end dates corresponding to the start and discontinuation dates of the newly initiated treatment. Secondary objectives assessed the severity of AD reported by clinical-assessed measures, including Eczema Area and Severity index (EASI) at baseline between the baricitinib and CS patient cohorts at W24. Clinician-assessed skin measures were reported in the eCRFs. Patient-reported outcome (PRO) measures were mapped to the related scheduled visit for the analysis of itch NRS and skin pain NRS over the 24-week period. For all objectives, data are also presented for other OS treatment cohorts. Statistical Analysis Treatment discontinuation was reported by estimates from Kaplan–Meier (KM) analysis of discontinuation rates from study commencement at baseline, and 95% confidence intervals (CI) were included to report median time to discontinuation. Patients lost to follow-up, including death or discontinuation from the study for any reason, were censored at the last available observation date. The reasons for discontinuation were presented as the number (n) and percentage (%) of patients who discontinued the study, as specified in the predefined reasons for discontinuation. Secondary outcomes are descriptively summarized by treatment cohorts (baricitinib and CS). EASI scores are reported as mean (standard deviation [SD]), while the ≤ 7-point change in EASI scores was reported as n (%) on the basis of patients with a combined reported outcome for body surface area (BSA) and itch NRS at baseline. For PROs, the itch NRS score and pain NRS score are presented as the n (%) of patients with a ≥ 4-point improvement. Missing data for primary outcomes was not imputed. For sensitivity analyses of secondary outcomes, single imputation approaches, including non-responder imputation (NRI) and last observation carried forward (LOCF), were used for selected outcomes. The NRI approach assumed that any patients with missing data were non-responders, regardless of the reason for missing data or potential treatment response in patients. In the case of LOCF, missing values for follow-up visits were replaced by the previously observed value for the patient. As a descriptive observational study, this analysis was not designed for formal statistical comparison between treatment cohorts, and consequently, any differences between cohorts cannot be interpreted as statistically significant. Results Kaplan–Meier Rate of Initial Patient Discontinuation and Reasons for Discontinuation Baseline characteristics have previously been reported [13]. Briefly, a total of 313 patients initiated treatment at baseline, with 87 in the baricitinib cohort and 113 patients receiving CS, respectively (Supplementary Table 1). Of the CS cohort, 1 received azathioprine, 82 received cyclosporine, 30 received methotrexate, and 0 received mycophenolate mofetil. In addition, 12 patients received systemic corticosteroids, 9 patients received other treatment, and 92 received other JAKi treatments (abrocitinib and upadacitinib), which became available after protocol approval. JAKi were used frequently in Germany (Supplementary Table 1; baricitinib = 42.5%; other JAKi = 42.4%), while CS were commonly used in France (30.1%), Spain (29.2%), and the UK (38.9%). At W24 (Fig. 1; Table 1), the KM cumulative incidence of treatment discontinuation for CS, baricitinib, and other JAKi was reported. The cumulative incidence of treatment discontinuation was reported as 50.8% (95% CI 40.9, 59.8) for CS, 29.0% (95% CI 19.8, 38.8) for baricitinib, and 18.8% (95% CI 11.5, 27.5) for other JAKi (Table 1). Kaplan–Meier plot for rate of discontinuation for conventional systemic, other JAKi, and baricitinib treatments at week W24 JAKi , Janus kinase inhibitor. Confidence intervals estimated by Kaplan–Meier method. The rate of discontinuation is represented by solid lines, with shaded areas indicating the 95% confidence intervals for each treatment cohort Table 1Overall patient population discontinuation at week 24Discontinuation up to week 24aBaricitinib (N = 87)Conventional systemics (N = 113)Other JAKi (N = 92)bKaplan–Meier cumulative incidence of initial treatment discontinuation, % (95% CI)c29.0 (19.8, 38.8)50.8 (40.9, 59.8)18.8 (11.5, 27.5)Patients with initial treatment discontinued, n (%)d25 (28.7)55 (48.7)17 (18.5)Primary lack of effectiveness7 (8.0)4 (3.5)6 (6.5)Secondary loss of effectiveness5 (5.7)2 (1.8)0 (0.0)Adverse events3 (3.4)13 (11.5)1 (1.1)Lack of compliance0 (0.0)1 (0.9)0 (0.0)Patient/physician decisione5 (5.7)9 (8.0)4 (4.3)AD remission0 (0.0)1 (0.9)0 (0.0)Treatment interruption ≥ 30 daysf2 (2.3)2 (1.8)3 (3.3)Data cutoff after interruptiong0 (0.0)1 (0.9)0 (0.0)Addition of a concomitant systemic treatmenth2 (2.3)17 (15.0)1 (1.1)Other1 (1.1)5 (4.4)2 (2.2)AD, atopic dermatitis; CI, Confidence interval; CRF, case report form; JAKi, Janus kinase inhibitor; N, total population in each cohort, n (%), number (percentage) of patients in the categoryaBaseline visit date is designated as study day 1. W24 corresponds to day 169bOther JAKi includes abrocitinib and upadacitinibcConfidence intervals estimated by Kaplan–Meier methoddSustained clinical response will be considered as a cause of treatment discontinuation in this analysiseIn accordance to CRF guidance, the use of “patient/physician decision” was selected when none of the other selections were applicable: a position, opinion, or judgment reached after consideration by a physician with reference to the patientfTreatment interruption included initiation, stopping, and re-initiation of the same systemic treatmentgCutoff was based on clinical protocols from phase III trials of systemic treatments in AD, which excluded patients exposed to systemic immunosuppressive/immunomodulating drugs within 4 weeks prior to baseline, consistent with treatment discontinuation [20, 34]hPatients were considered as non-responders from the point in time concomitant systemic treatment commenced and were censored Overall patient population discontinuation at week 24 AD, atopic dermatitis; CI, Confidence interval; CRF, case report form; JAKi, Janus kinase inhibitor; N, total population in each cohort, n (%), number (percentage) of patients in the category aBaseline visit date is designated as study day 1. W24 corresponds to day 169 bOther JAKi includes abrocitinib and upadacitinib cConfidence intervals estimated by Kaplan–Meier method dSustained clinical response will be considered as a cause of treatment discontinuation in this analysis eIn accordance to CRF guidance, the use of “patient/physician decision” was selected when none of the other selections were applicable: a position, opinion, or judgment reached after consideration by a physician with reference to the patient fTreatment interruption included initiation, stopping, and re-initiation of the same systemic treatment gCutoff was based on clinical protocols from phase III trials of systemic treatments in AD, which excluded patients exposed to systemic immunosuppressive/immunomodulating drugs within 4 weeks prior to baseline, consistent with treatment discontinuation [20, 34] hPatients were considered as non-responders from the point in time concomitant systemic treatment commenced and were censored In the baricitinib cohort, patients primarily discontinued owing to primary lack of effectiveness (n = 7; 8.0%), followed by patient or physician decision, and secondary lack of effectiveness (n = 5; 5.7% for both), with other JAKi reporting a similar trend in discontinuation due to primary lack of effectiveness (n = 6, 6.5%) or patient or physician decision (n = 4, 4.3%). In patients who received CS, patients were reported to have discontinued initial treatment owing to the addition of concomitant systemic immunomodulatory treatment (n = 17, 15.0%), adverse events (n = 13, 11.5%), patient or physician decision (n = 9, 8.0%), and primarily lack of effectiveness (n = 4, 3.5%). Results for other individual treatment cohorts are reported in Supplementary Table 2. Physician and Patient-Reported Outcomes at Week 24 Table 2 describes physician-reported clinical outcomes for EASI. At baseline, patients reported a mean (SD) EASI total score of 19.8 (10.8) for CS, 15.8 (10.6) for baricitinib, and 16.4 (9.2) for other JAKi. At W24, the mean (SD) EASI total score for CS was 10.1 (11.0), 5.2 (6.1) for baricitinib, and 5.3 (8.7) for other JAKi. Overall, 23.0% of the CS, 41.2% of the baricitinib, and 62.1% of the other JAKi cohorts achieved an EASI75 at W24. Patients with an EASI ≤ 7 score at W24 (Table 3) were reported as 81.0% for the baricitinib cohort, 84.2% for the other JAKi, and 63.1% for CS. Table 2EASI outcomes for the overall population at week 24 from baselineEASI scoreBaricitinib (N = 87)Conventional systemics (N = 113)Other JAKi (N = 92)aEASI total scoreBaseline, n809883Mean (SD)15.8 (10.6)19.8 (10.8)16.4 (9.2)Week 24b, n687470Mean (SD)5.2 (6.1)10.1 (11.0)5.3 (8.7)EASI 75cWeek 24, n687466n (%)d28 (41.2)17 (23.0)41 (62.1)EASI, Eczema area and Severity Index; EASI75, a ≥ 75% improvement in EASI from baseline; JAKi, Janus kinase inhibitor; LOCF, last observation carried forward; N, total population; n, number of patients in each cohort; NRI, non-responder imputation; SD, standard deviation; %, percentage of patientsaOther JAKi includes abrocitinib and upadacitinibbMissing data were imputed using LOCF imputation for week-24 outcomescMissing data were imputed using NRI imputation for week-24 outcomesdThe percentage of patients is based on the number of patients with a response and available data for EASI at week 24 EASI outcomes for the overall population at week 24 from baseline EASI, Eczema area and Severity Index; EASI75, a ≥ 75% improvement in EASI from baseline; JAKi, Janus kinase inhibitor; LOCF, last observation carried forward; N, total population; n, number of patients in each cohort; NRI, non-responder imputation; SD, standard deviation; %, percentage of patients aOther JAKi includes abrocitinib and upadacitinib bMissing data were imputed using LOCF imputation for week-24 outcomes cMissing data were imputed using NRI imputation for week-24 outcomes dThe percentage of patients is based on the number of patients with a response and available data for EASI at week 24 Table 3Proportion of patients with EASI ≤ 7 at baseline and week 24, overall and by BSA and Itch NRS subgroupsEASI ≤ 7aBaricitinib (N = 87)Conventional systemics (N = 113)Other JAKi (N = 92)bBaselineOverall, nxc648464n (%)d14 (21.9)11 (13.1)11 (17.2)BSA and Itch NRS subgroupsBSA ≤ 40% and Itch NRS ≥ 7, nx302327n (%)6 (20.0)4 (17.4)4 (14.8)BSA ≤ 40% and Itch NRS < 7, nx272519n (%)8 (29.6)6 (24.0)7 (36.8)BSA > 40% and Itch NRS ≥ 7, nx4220n (%)0 (0.0)0 (0.0)0 (0.0)BSA > 40% and Itch NRS < 7, nx3140n (%)0 (0.0)1 (7.1)0 (0.0)Week 24Overall, nxe586557n (%)f47 (81.0)41 (63.1)48 (84.2)BSA and Itch NRS subgroupsBSA ≤ 40% and Itch NRS ≥ 7, nx251925n (%)20 (80.0)15 (78.9)21 (84.0)BSA ≤ 40% and Itch NRS < 7, nx251616n (%)20 (80.0)10 (62.5)15 (93.8)BSA > 40% and Itch NRS ≥ 7, nx41811n (%)3 (75.0)12 (66.7)7 (63.6)BSA > 40% and Itch NRS < 7, nx4125n (%)4 (100.0)4 (33.3)5 (100.0)BSA, body surface area; EASI, Eczema area and Severity Index; JAKi, Janus kinase inhibitor; LOCF, last observation carried forward; N, total population; n, number of patients in each cohort; nx, number of patients with this outcome at baseline; NRI, non-responder imputation; NRS, numerical rating scale; %, percentage of patientsaMissing data were imputed using LOCF imputation for week-24 outcomesbOther JAKi includes abrocitinib and upadacitinibcOverall nx includes patients with evaluable EASI data regardless of available BSA or itch NRS data. Patients with missing BSA and itch NRS are includeddThe percentage of patients at baseline is calculated as number of patients with a response and available outcome data at baselineeSubgroup nx is restricted to patients with evaluable EASI, BSA, and itch NRS data. Patients with missing BSA or itch NRS data are excluded from the subgroupfThe percentage of patients at week 24 is calculated as number of patients with a response and available outcome data at week 24 Proportion of patients with EASI ≤ 7 at baseline and week 24, overall and by BSA and Itch NRS subgroups BSA, body surface area; EASI, Eczema area and Severity Index; JAKi, Janus kinase inhibitor; LOCF, last observation carried forward; N, total population; n, number of patients in each cohort; nx, number of patients with this outcome at baseline; NRI, non-responder imputation; NRS, numerical rating scale; %, percentage of patients aMissing data were imputed using LOCF imputation for week-24 outcomes bOther JAKi includes abrocitinib and upadacitinib cOverall nx includes patients with evaluable EASI data regardless of available BSA or itch NRS data. Patients with missing BSA and itch NRS are included dThe percentage of patients at baseline is calculated as number of patients with a response and available outcome data at baseline eSubgroup nx is restricted to patients with evaluable EASI, BSA, and itch NRS data. Patients with missing BSA or itch NRS data are excluded from the subgroup fThe percentage of patients at week 24 is calculated as number of patients with a response and available outcome data at week 24 Next, the Itch NRS and skin Pain NRS were reported at W24 from baseline (Fig. 2 and Supplementary Table 4). Overall population patient-reported outcomes for Itch NRS, and Pain NRS. JAKi, Janus-kinase inhibitor; N, total population; NRI, non-responder imputation; NRS, numerical rating scale. Missing data were imputed using NRI At baseline, an Itch NRS ≥ 4 was reported for 80.8% of patients who received baricitinib, 85.1% of the other JAKi cohort, and 87.1% of the CS cohort. After 24 weeks of treatment, 42.0% of patients treated with baricitinib reported a ≥ 4-point improvement in itch, similar to the 38.5% of the other JAKi cohort. For the CS cohort, 14.3% reported this outcome. Patients with a pain NRS score ≥ 4 at baseline were reported as 60.9% for the baricitinib cohort, 56.5% for the other JAKi cohort, and 59.3% for CS cohort. At W24, a ≥ 4-point improvement in Pain NRS was reported in 38.1% of the baricitinib cohort, 38.2% of the other JAKi, but only 16.7% of the CS cohort. Discussion The emergence of novel therapeutic options has expanded the list of available OS therapeutics for the treatment and management of moderate-to-severe AD [10]. While clinical guidelines recommend their use, real-world data on current treatment practices and the effectiveness of these therapies remain limited [10, 14]. To address this gap, the AD-REAL analysis reports on the pattern of treatment discontinuation and the effectiveness of OS therapies in real-world practice across four European countries. At week 24, the Kaplan–Meier cumulative incidence of initial treatment discontinuation was numerically lower for baricitinib and other JAKi than CS. The reasons for treatment discontinuations were different between the cohorts. The primary reasons for patient discontinuation in the CS were the occurrence of adverse events and the addition of concomitant immunomodulatory drug, while the baricitinib and other JAKi cohorts primarily discontinued owing to lack of treatment effectiveness. At baseline, 70.1% of patients in the baricitinib cohort were reported to have previous exposure to oral and/or biologic systemic treatments for AD (similar to the other JAKi), while 31.0% of the CS cohort reported this outcome [13]. Additionally, of the patients in the baricitinib cohort, 44.8% reported previous failure in systemic therapy treatment due to primary lack of effectiveness or secondary loss of effectiveness, while 10.6% of the CS cohort reported this outcome. These characteristics may indicate that at baseline, the baricitinib and other JAKi cohorts experienced more difficult-to-manage AD. One explanation for these differences may be that CS, including cyclosporine, are used as first-line therapies for the treatment of moderate-to-severe AD, while the use of JAKi may be recommended as second-line treatment following an inadequate response to CS [10, 15]. Therefore, these baseline differences may reflect the real-world design of this study and current treatment practices for AD in Europe. These results follow a similar trend to other studies where patients who received CS therapies had a higher incidence of treatment discontinuation compared to dupilumab, with a higher frequency of adverse events [16]. While conventional agents such as cyclosporin are routinely used in daily practice for rapid relief of severe AD and to bridge the waiting period for slower-acting therapies, these treatments are often discontinued within the first year to avoid adverse events [10, 17]. Data from the AD-REAL study follow a similar trend, as a high frequency of patients suspended CS therapy (11.5%) owing to adverse events, while both the baricitinib and the other JAKi cohorts had numerically lower discontinuation rates for this outcome. In RCTs, relative response outcomes, such as the percentage of patients achieving ≥ 75% improvement from baseline in EASI, are commonly used to assess the severity of AD over time [18–21]. However, relative outcomes may be limited as they depend on baseline disease severity. As a result, in clinical practice, absolute disease severity measurements, such as EASI ≤ 7 or an Itch NRS ≤ 4, may serve as an alternative treatment goal at 6 months, especially in the instance of moderate AD populations such as those enrolled in the AD-REAL study [22]. After 24 weeks of treatment, it was found that 81.0% of patients treated with baricitinib similar to other JAKis (84.2%) achieved an EASI ≤ 7 score. This result is comparable to other clinical studies, where nearly 70% of patients reported a similar EASI score after 6 months of treatment with baricitinib, which was maintained up to 12 months [23]. Additionally, this study reported improved QoL outcomes in patients who received baricitinib and other JAKi. AD is known to carry a substantial burden to patients and negatively impacts their QoL [24–28]. Severe and chronic itch represents a significant burden to patients, given its association with sleep disturbances and overall impairment to patient QoL [29]. This study found that baricitinib, and other JAKi, were effective in improving PROs, including itch and pain scores, while CS reported numerically lower improvement in these outcomes. These findings for baricitinib are consistent with previous efficacy reports from the BREEZE-AD trials and reinforce the efficacy in patients with lower BSA but disease burden related to itch [30–33]. Together, these findings help to inform on the current treatment discontinuation patterns and therapeutic effectiveness of CS, baricitinib, and other classes of JAKi in real-world clinical practice. This study has several limitations. Owing to the nature of real-world studies, the decision to discontinue treatment was at the discretion of the patient and physician, which may be influenced by more than one factor. Secondly, differences at baseline were not controlled for in this study. Doses and uptitrations were collected for baricitinib but not for abrocitinib and upadacitinib. This makes numerical differences in discontinuation rates relating to efficacy and safety difficult to interpret. Additionally, patients lost to follow-up and varying rates of study participation can introduce attrition bias. This study was conducted during the COVID-19 pandemic, which impacted the sample size recruited to the study and may have led to higher rates of attrition, particularly regarding PROs. While single imputation approaches, including NRI and LOCF, were used to summarize post-baseline clinical measures and PRO endpoints, the varying rates of attrition and missing data may have still introduced unpredictable bias to the study and impacted the interpretation of results and comparability of data between the cohorts. Therefore, given the descriptive nature of this study, any direct comparison between cohorts should be interpreted with caution, owing to the absence of a formal statistical powering, high rate of attrition, and the exploratory nature of the other JAKi cohort. Conclusions Findings from the descriptive AD-REAL observational cohort study suggest that, in routine clinical practice, patients treated with baricitinib and other JAKi had numerically lower instances of treatment discontinuation and numerically improved clinical outcomes after 24 weeks of treatment than the conventional OS cohort. Supplementary Information Below is the link to the electronic supplementary material. Supplementary file1 (PDF 101 KB)