Real-World Use of Upadacitinib for Moderate-to-Severe Atopic Dermatitis in Saudi Arabia: A Single-Centre Retrospective Descriptive Analysis Authors: Hamad Almutlaq Published: 2026 ABSTRACT AbstractBackgroundUpadacitinib is an oral Janus kinase 1 inhibitor for moderate-to-severe atopic dermatitis, although real-world evidence from Middle Eastern clinical practice remains limited. The objective of the present study was to describe treatment patterns, physician-documented response, symptom improvement, adverse events, discontinuation, and laboratory monitoring in patients treated with upadacitinib for moderate-to-severe atopic dermatitis.MethodsThis retrospective descriptive analysis included 30 patients treated with upadacitinib at King Fahd University Hospital between 2023 and 2025. Data were extracted from clinical records and summarized descriptively. Outcomes included dose, treatment duration, response timing, itch and sleep improvement, flare reduction, adverse events, laboratory abnormalities, and discontinuation.ResultsClinical response was evaluable in 28 patients; two were not evaluable because follow-up documentation was insufficient or unclear. Qualitative physician-documented improvement was recorded in 27 of the 28 evaluable patients, and response was documented within four weeks in 23 of 28. Among patients with paired symptom data, the median pruritus NRS decreased from 8.0 to 2.0, the median sleep-disturbance rating decreased from 8.0 to 2.0, and monthly flare frequency decreased from 5.0 to 1.0. A reduction of at least four points in pruritus NRS occurred in 25 of 28 patients. Retained adverse events occurred in 13 patients. Five patients discontinued treatment, including two because of safety, infection, or laboratory concerns.ConclusionFrequent physician-documented improvement was observed in patients with evaluable follow-up; however, five patients discontinued upadacitinib, including two due to safety, infection, or laboratory concerns. These descriptive findings suggest careful interpretation, patient selection, baseline assessment and frequent clinical and laboratory follow-up.Keywords: dermatitis, atopic, Janus kinase inhibitors, Upadacitinib Introduction Atopic dermatitis (AD) is a chronic inflammatory skin disorder that affects about one-fifth of children and as many as 7% of adults. The hallmark of AD is intense pruritus; patients frequently experience sleep disturbances and psychiatric comorbidities, and the disease has been identified as one of the skin conditions with the greatest impact on quality‑adjusted life expectancy.1 Severity is often defined using the Scoring Atopic Dermatitis (SCORAD) scale: SCORAD < 25 indicates mild disease, 25–50 indicates moderate disease and > 50 indicates severe disease.2 Despite advances in treatment, a considerable proportion of patients with severe AD remain undertreated; undertreatment is driven by the underestimation of AD’s economic, social and humanistic impact and by inconsistent access to effective therapies. Patients and families bear a complex burden that includes itch-associated sleep disturbance, functional impairment, comorbid allergic or atopic disorders, neuropsychiatric sequelae and significant out-of-pocket costs.2 In the last 10 years, the treatment options for moderate to severe AD have grown significantly. For many years, systemic immunosuppressants like cyclosporine and methotrexate were used off label with modest efficacy and significant toxicity. In more recent times, biologic drugs like dupilumab have provided targeted inhibition of the interleukin 4 and 13 pathways, while Janus kinase (JAK) inhibitors have proven to be a promising oral alternative. The American Academy of Dermatology (AAD) and the American Academy of Allergy, Asthma and Immunology (AAAAI) conditionally recommend the use of oral JAK inhibitors (abrocitinib and upadacitinib) for moderate to severe AD that is resistant to topical treatments or biologics.3 The guidelines emphasize that JAK inhibitors are highly effective and have a quick onset of action in Phase III trials, but warn that the long-term safety of these drugs is uncertain, which limits strong recommendations. Upadacitinib is an oral, selective Janus kinase 1 inhibitor approved for moderate-to-severe atopic dermatitis (AD), offering rapid and clinically meaningful improvement in disease activity. Its use nevertheless requires careful benefit–risk assessment. Current United States prescribing information carries boxed warnings concerning serious infections, mortality, malignancy, major adverse cardiovascular events (MACE), and thrombosis, although several comparative warnings originate from a safety trial of another JAK inhibitor in an older, cardiovascular-risk-enriched rheumatoid arthritis population.4 European regulatory guidance similarly recommends that JAK inhibitors be used in patients aged ≥65 years, current or former long-term smokers, and patients with cardiovascular or malignancy risk only when suitable alternatives are unavailable, with additional caution in patients with venous thromboembolism risk factors.5 Recent evidences have helped contextualise, but has not eliminated, these concerns. An integrated upadacitinib analysis included 8632 patients and 27,164.2 patient-years across several inflammatory diseases,6 while an AD-specific analysis comprising 2683 patients and 9187.7 patient-years reported low rates of malignancy, MACE, and venous thromboembolism over up to six years.7 However, the AD population in the broader analysis overlapped with the AD-specific trial programme, and older or otherwise high-risk patients remained comparatively under-represented. Real-world evidence on upadacitinib in atopic dermatitis is expanding internationally, but region-specific data on tolerability, laboratory monitoring findings and physician-documented outcomes in Middle Eastern and Saudi patients remain limited. Furthermore, real-world analyses help clinicians and policymakers understand how novel therapies perform outside of clinical trials, including in patients with prior biologic exposure or comorbidities. Therefore, the present study aimed to describe the real-world use, tolerability, laboratory monitoring findings and physician-documented clinical outcomes of upadacitinib in adolescents and adults with moderate-to-severe atopic dermatitis. Methodology Study Design and Setting This study was conducted as a retrospective descriptive analysis of patients with moderate-to-severe atopic dermatitis who received upadacitinib as part of routine clinical care at King Fahd University Hospital. The study covered patients treated between 2023 and 2025, with an analysis cut-off date of 20 May 2025. The analysis was designed to describe treatment patterns, treatment duration, physician-documented clinical response, symptom improvement, adverse events, treatment discontinuation, and laboratory monitoring. No comparative effectiveness analysis, hypothesis testing, regression modelling, survival analysis, or causal inference was planned because of the descriptive study design and limited sample size. Ethical Approval This retrospective descriptive analysis was approved by the Institutional Review Board (IRB) of Imam Abdulrahman Bin Faisal University, Saudi Arabia, NCBE Registration No. HAP-05-D-003 (IRB approval number: IRB-2025-01-0805; approval date: 13 November 2025). The retrospective data extraction and analysis were initiated only after obtaining IRB approval. The study involved retrospective review of existing medical records with no direct patient contact and the use of de-identified data, the informed consent was waived by the IRB. The study complies with Declaration of Helsinki. Study Population Patients were included if they had a documented diagnosis of atopic dermatitis, were clinically classified as having moderate or severe disease at baseline, and received upadacitinib during the study period. Patients were included in the full cohort if information on baseline diagnosis, baseline disease severity, upadacitinib dose, treatment start date, and treatment status at the analysis cut-off was available. Patients with incomplete or unclear follow-up documentation were retained in the full cohort for baseline, treatment exposure, safety, and laboratory summaries where data were available. However, they were classified as not evaluable for clinical response and symptom-response outcomes when response could not be reliably assessed from the available documentation. Treatment Selection and Concomitant Therapy Upadacitinib was prescribed by the treating clinicians as part of routine clinical care before inclusion in this retrospective analysis, and treatment allocation was neither randomized nor governed by a study protocol. The medical records did not consistently capture all reasons for selecting upadacitinib rather than another systemic treatment. The type, timing, dose, and duration of concomitant topical or systemic atopic dermatitis treatments during follow-up were also not consistently available. Consequently, the determinants of treatment selection and potential confounding by concomitant therapy could not be formally evaluated. Sample-Size Considerations No formal a priori sample-size or statistical power calculation was performed because this was a retrospective, non-comparative descriptive analysis and no hypothesis testing or comparative effect estimation was planned. The final sample consisted of 30 patients who met the eligibility and data-availability criteria during the defined study period at the participating centre. The available sample was considered suitable for describing local treatment patterns and documented clinical experience but was not expected to provide precise estimates of uncommon outcomes, support adequately powered subgroup comparisons, or establish comparative effectiveness or safety. Data Source and Variables Data were extracted retrospectively from available clinical records and entered into a structured dataset. The extracted variables included study identification code, diagnosis, baseline disease severity, comorbidities, previous treatment exposure, upadacitinib dose, treatment start date, treatment status at the analysis cut-off, treatment duration, discontinuation status, reason for discontinuation, physician-documented clinical response, timing of first documented clinical response, itch improvement, sleep improvement, flare reduction, baseline and follow-up flare frequency, adverse events, adverse-event severity, adverse-event management, adverse-event resolution, laboratory monitoring status, number of laboratory-monitoring visits, and baseline and follow-up laboratory parameters. Previous treatment exposure and comorbidities were recorded as documented in the clinical records. Because individual patients could have more than one comorbidity or previous treatment, these variables were treated as non-mutually exclusive categories. Directly recorded baseline and follow-up 0–10 pruritus numerical rating scale scores and 0–10 sleep-disturbance ratings were also extracted where available. BMI and smoking status were not consistently available in the retrospective records and were therefore not included in the analysis. Treatment Exposure and Treatment Duration Upadacitinib exposure was summarized according to prescribed dose, categorized as 15 mg or 30 mg. Treatment status at the analysis cut-off was categorized as ongoing or stopped. Treatment duration was calculated in months from the upadacitinib start date to the relevant end point. For patients who were continuing treatment, the end point was the analysis cut-off date, 20 May 2025. For patients who discontinued treatment, the end point was the documented discontinuation date or the recorded treatment end point available in the dataset. Treatment duration was summarized for the full cohort, for patients continuing treatment, and for patients who discontinued treatment. Reasons for discontinuation were grouped descriptively into safety/infection or laboratory concern, ineffectiveness or no improvement, medication access or lost follow-up, and other or unclear reasons after data cleaning. Physician-Documented Clinical Response Clinical response was assessed from physician-documented clinical notes and categorized according to the response descriptions available in the dataset. Response categories included dramatic improvement, major improvement, moderate improvement, partial improvement, no meaningful improvement, and not evaluable. Patients with adequate follow-up documentation were included in the evaluable response denominator. Patients with insufficient or unclear follow-up documentation were retained in the full cohort but excluded from evaluable response summaries. The timing of first documented clinical response was recorded in weeks when available. Response timing was categorized as rapid response when documented within 2 weeks, early response when documented within 3–4 weeks, delayed or partial response when documented within 5–8 weeks, no adequate response by 16 weeks, or not evaluable. These response outcomes were treated as physician-documented clinical observations and not as validated objective efficacy outcomes because repeated validated measures such as EASI, SCORAD, or DLQI was not available. Clinical response and symptom outcomes were extracted from routine follow-up notes recorded by the treating clinicians. Because these clinicians were directly involved in patient care, they were aware that the patients were receiving upadacitinib and were not independent blinded outcome assessors. The retrospective records did not permit confirmation that the same clinician performed all follow-up assessments. No independent reassessment or outcome adjudication was undertaken. Consequently, these outcomes were interpreted as physician-documented clinical observations rather than standardized or independently verified efficacy outcomes. Symptom Improvement and Flare Reduction Directly recorded 0–10 pruritus numerical rating scale scores, 0–10 sleep-disturbance ratings, and monthly flare frequency were summarized at baseline and at the documented follow-up assessment. Among patients with paired observations, absolute reduction was calculated as the baseline value minus the follow-up value. For pruritus, the proportions of patients with a reduction of at least four points and with a follow-up score of 0 or 1 were also calculated. The exact validated instrument used for the sleep-disturbance rating and a uniform follow-up assessment time were not available; therefore, sleep-score changes were treated as descriptive routine-care observations. Narrative clinical descriptions were separately mapped by the research team into qualitative categories comprising dramatic, major, moderate, partial, or no meaningful improvement. These qualitative categories were not derived from PGA, EASI, SCORAD, or another validated instrument, and no quantitative thresholds were applied. Adverse Events and Discontinuation-Related Safety Outcomes Adverse events were summarized descriptively for the full cohort. Recorded adverse events included infection-related events, herpes zoster, herpes simplex or eczema herpeticum, acne, weight gain, neutropenia, and documented laboratory abnormalities. Because patients could experience more than one adverse event, adverse-event categories were treated as non-mutually exclusive. Adverse-event severity was summarized using the documented severity categories in the dataset. Where available, adverse-event management and resolution were reviewed descriptively. Therefore, the final adverse-event analysis was based only on retained adverse events after data cleaning. Patients who discontinued upadacitinib were described separately according to study identification code, dose, treatment duration, baseline severity, reason for stopping, discontinuation category, physician-documented response, and whether a retained adverse event was documented. Laboratory Monitoring and Laboratory Parameters Laboratory monitoring was summarized according to whether post-baseline laboratory monitoring was documented. Among monitored patients, the number of laboratory-monitoring visits was summarized descriptively. Laboratory parameters included absolute neutrophil count, alanine aminotransferase, aspartate aminotransferase, low-density lipoprotein cholesterol, triglycerides, bilirubin, and lactate dehydrogenase. For absolute neutrophil count, baseline and nadir values were summarized. For alanine aminotransferase, aspartate aminotransferase, low-density lipoprotein cholesterol, triglycerides, bilirubin, and lactate dehydrogenase, baseline and peak values were summarized. Laboratory abnormalities were reported as documented in the dataset and summarized using the available denominators. Statistical Analysis All analyses were descriptive. Categorical variables were summarized as frequencies and percentages. Percentages were calculated using the relevant denominator for each outcome, including the full cohort, evaluable patients, monitored patients, paired observations, or discontinued patients, as appropriate. Paired baseline and follow-up pruritus scores, sleep-disturbance ratings, and monthly flare frequency were summarized using medians, interquartile ranges, ranges, and absolute within-patient reductions. Pruritus responder outcomes were summarized as frequencies and percentages. Continuous variables were summarized using median, interquartile range, and range because of the small sample size and the likely non-normal distribution of several clinical and laboratory variables. These measures were considered more appropriate than the mean and standard deviation because they are less influenced by skewed distributions and extreme values. This included treatment duration, timing of first documented clinical response, flare frequency, number of laboratory-monitoring visits, and laboratory parameters. No hypothesis testing, regression analysis, survival analysis, or comparative subgroup analysis was performed. Missing or insufficiently documented outcomes were not imputed. Patients with insufficient response documentation were classified as not evaluable for response-related analyses but were retained in the full cohort denominator where applicable. Results The analysis addressed the principal descriptive questions of how upadacitinib was used in routine practice, the timing and extent of physician-documented clinical and symptom improvement, treatment duration and discontinuation, and the occurrence and management of adverse events and laboratory abnormalities. The study included 30 patients with atopic dermatitis. The cohort was a young median age, with a slight male predominance. Baseline disease severity was almost evenly distributed between moderate and severe disease, although severe disease was slightly more frequent. Nearly half of the patients had documented comorbidities, with asthma and dyslipidemia being the most commonly recorded. Previous advanced systemic therapy was documented in a small number of patients: four had received dupilumab, one had received baricitinib, and one had received another biologic. Because these subgroups were small, treatment outcomes were not analysed according to previous biologic or JAK-inhibitor exposure. Documented previous topical therapy was recorded in 15 patients; however, incomplete retrospective capture of earlier topical treatment cannot be excluded. (Table 1). Table 1Baseline Clinical Characteristics and Previous Treatment Exposure (n=30)Characteristicn (%)Age, in years25.0 (20.0–33.3)Males16 (53.3%)Diagnosis: atopic dermatitis30 (100.0)Baseline severity: moderate14 (46.7)Baseline severity: severe16 (53.3)Any documented comorbidity14 (46.7)No documented comorbidity15 (50.0)Comorbidity status unknown1 (3.3)Asthma8 (26.7)Dyslipidemia5 (16.7)Diabetes mellitus2 (6.7)Obesity2 (6.7)Allergic rhinitis1 (3.3)Psychiatric comorbidity1 (3.3)Alopecia areata/universalis2 (6.7)Hepatic steatosis/NASH1 (3.3)Ethnic neutropenia at baseline1 (3.3)Ectodermal dysplasia1 (3.3)Documented Previous topical therapy15 (50.0)Previous oral corticosteroid5 (16.7)Previous phototherapy4 (13.3)Previous cyclosporine4 (13.3)Documented previous dupilumab exposure4 (13.3)Previous methotrexate1 (3.3)Previous azathioprine1 (3.3)Previous mycophenolate mofetil1 (3.3)Previous omalizumab1 (3.3)Documented previous Baricitinib exposure1 (3.3)Documented previous another biologic exposure, unspecified1 (3.3)Notes: Patients may have more than one comorbidity and more than one previous treatment; therefore, category totals may exceed 100%. Previous treatment variables reflect treatments documented in the available medical records. Absence of documentation should not be interpreted as confirmation that a treatment had never been used.Abbreviations: AD, atopic dermatitis; NASH, non-alcoholic steatohepatitis. Baseline Clinical Characteristics and Previous Treatment Exposure (n=30) Notes: Patients may have more than one comorbidity and more than one previous treatment; therefore, category totals may exceed 100%. Previous treatment variables reflect treatments documented in the available medical records. Absence of documentation should not be interpreted as confirmation that a treatment had never been used. Abbreviations: AD, atopic dermatitis; NASH, non-alcoholic steatohepatitis. Upadacitinib was prescribed at both 15 mg and 30 mg doses, with a modest predominance of the 15 mg dose. Most patients remained on treatment at the analysis cut-off, and the overall median treatment duration was 12.0 months. Treatment discontinuation was documented in a small proportion of patients, with safety, infection, or laboratory concerns being the most frequent category among those who stopped therapy (Table 2). Table 2Upadacitinib Exposure, Treatment Duration, and DiscontinuationVariableResultUpadacitinib dose: 15 mg16 (53.3)Upadacitinib dose: 30 mg14 (46.7)Treatment ongoing at analysis cut-off25 (83.3)Treatment stopped by analysis cut-off5 (16.7)Planned future discontinuation while still ongoing1 (3.3)Treatment duration, months, median (IQR) [range]12.0 (9.2–18.3) [2.8–23.6]Treatment duration among ongoing patients, months, median (IQR) [range]13.6 (10.7–20.8) [2.8–23.6]Treatment duration among stopped patients, months, median (IQR) [range]8.0 (7.8–12.0) [6.0–12.0]Reason for stopping: safety/infection or laboratory concern2 (40.0)Reason for stopping: ineffectiveness/no improvement1 (20.0)Reason for stopping: medication access/lost follow-up1 (20.0)Reason for stopping: other/unclear after data cleaning1 (20.0)Note: Percentages for stopping reasons are calculated using the five discontinued patients as the denominator.Abbreviation: IQR, interquartile range. Upadacitinib Exposure, Treatment Duration, and Discontinuation Note: Percentages for stopping reasons are calculated using the five discontinued patients as the denominator. Abbreviation: IQR, interquartile range. Physician-documented improvement was observed in most evaluable patients, with the majority classified as having either major or dramatic improvement. However, these findings should be interpreted as documented clinical response rather than confirmed objective efficacy, as repeated validated severity scores were not available (Table 3). Table 3Qualitative Physician-Documented Clinical Response and Categorized Timing of ResponseOutcomen (%) or Summary StatisticDramatic improvement8 (26.7)Major improvement, including one patient before discontinuation15 (50.0)Moderate improvement1 (3.3)Partial improvement3 (10.0)No meaningful improvement1 (3.3)Not evaluable2 (6.7)Any physician-documented improvement among evaluable patients27 (96.4)Rapid response (<=2 weeks)7 (23.3)Early response (3–4 weeks)16 (53.3)Delayed/partial response (5–8 weeks)4 (13.3)No adequate response by 16 weeks1 (3.3)Not evaluable2 (6.7)Notes: Narrative response descriptions were mapped by the research team into qualitative categories. These categories were not based on PGA, IGA, EASI, SCORAD, or predefined percentage-improvement thresholds and should not be interpreted as standardized efficacy endpoints.Abbreviation: IQR, interquartile range. Qualitative Physician-Documented Clinical Response and Categorized Timing of Response Notes: Narrative response descriptions were mapped by the research team into qualitative categories. These categories were not based on PGA, IGA, EASI, SCORAD, or predefined percentage-improvement thresholds and should not be interpreted as standardized efficacy endpoints. Abbreviation: IQR, interquartile range. Documented symptom outcomes were consistent with the physician-recorded response pattern. Most evaluable patients experienced improvement in itch and sleep disturbance, along with a reduction in disease flares. The recorded flare frequency decreased from baseline to follow-up, supporting a documented reduction in flare burden during treatment (Table 4). Table 4Paired Changes in Directly Recorded Pruritus, Sleep-Disturbance, and Flare-Frequency OutcomesOutcomePaired nBaselineMedian (IQR) [Range]Follow-UpMedian (IQR) [Range]MedianReduction (IQR) [Range]Pruritus NRS, 0–10288.0 (7.8–9.0) [7–9]2.0 (1.0–2.0) [1–8]6.0 (5.0–8.0) [0–8]Sleep-disturbance rating, 0–10288.0 (7.5–8.5) [6–10]2.0 (0.0–2.0) [0–8]6.0 (4.0–8.0) [0–10]Flares per month285.0 (4.8–7.0) [4–8]1.0 (0.8–1.0) [0–6]4.0 (3.0–6.3) [0–8]Responder outcomen/N (%)≥4-point reduction in pruritus NRS25/28 (89.3%)Follow-up pruritus NRS 0/18/28 (28.6%)Any reduction in sleep-disturbance rating27/28 (96.4%)Any reduction in monthly flare frequency27/28 (96.4%)No flares at documented follow-up7/28 (25.0%)Notes: Pruritus NRS and sleep-disturbance ratings were directly recorded in routine medical records. The exact sleep instrument and a uniform follow-up assessment time were unavailable; therefore, sleep findings are descriptive. Reduction was calculated as baseline minus follow-up. Paired Changes in Directly Recorded Pruritus, Sleep-Disturbance, and Flare-Frequency Outcomes Notes: Pruritus NRS and sleep-disturbance ratings were directly recorded in routine medical records. The exact sleep instrument and a uniform follow-up assessment time were unavailable; therefore, sleep findings are descriptive. Reduction was calculated as baseline minus follow-up. Retained adverse events were documented in fewer than half of the cohort. Infection-related adverse events, including herpes zoster and herpes simplex/eczema herpeticum, represented an important safety finding. Acne and laboratory abnormalities were also documented. Most adverse events were mild or moderate, and no severe adverse event was recorded in the dataset. Upadacitinib was continued with monitoring in most patients with adverse events, while temporary interruption or discontinuation was used in selected cases based on clinical context (Table 5). Table 5Adverse Events and Laboratory AbnormalitiesDomainOutcomePatients, n/N (%)AE Records, nClinical Details, Management, or OutcomeOverall adverse-event burdenAny retained adverse event13/30 (43.3)18Adverse events were documented in 13 unique patients.Overall adverse-event burdenNo retained adverse event17/30 (56.7)—No adverse event was retained after data cleaning.Infection-related adverse eventsAny infection-related adverse event6/30 (20.0)7Included recurrent bacterial/skin infections, herpes zoster, herpes simplex, and eczema herpeticum.Infection-related adverse eventsHerpes zoster2/30 (6.7)2Managed with antiviral therapy; treatment was temporarily held or discontinued according to clinical context.Infection-related adverse eventsHerpes simplex/eczema herpeticum3/30 (10.0)4Episodes were managed with antiviral therapy; documented outcomes were resolved after treatment or interruption.Infection-related adverse eventsRecurrent bacterial/skin infections1/30 (3.3)1Managed with oral antibiotics and infection surveillance; treatment was discontinued because infection risk coexisted with neutropenia.Dermatologic adverse eventsAcne4/30 (13.3)4Managed with topical anti-acne therapy; upadacitinib was continued with monitoring.Metabolic/weight-related adverse eventsWeight gain1/30 (3.3)1Managed with diet and activity counselling; the event remained ongoing but non-serious.Laboratory abnormalitiesAny documented laboratory abnormality5/30 (16.7)6Included neutropenia, lipid elevation, liver enzyme elevation, bilirubin/GGT elevation, and LDH elevation.Laboratory abnormalitiesNeutropenia3/30 (10.0)3Managed with repeat complete blood count monitoring; treatment was continued, interrupted, or discontinued according to coexisting infection risk.Laboratory abnormalitiesLipid elevation2/30 (6.7)2Managed with lifestyle advice, repeated lipid monitoring, and clinical follow-up.Laboratory abnormalitiesLiver enzyme/bilirubin/LDH elevation2/30 (6.7)2Managed with repeat laboratory monitoring and evaluation for alternative causes or drug-induced liver injury.Severity among patients with adverse eventsMild events only6/13 (46.2)—No moderate event was documented in these patients.Severity among patients with adverse eventsModerate events only3/13 (23.1)—No severe adverse event was documented in the dataset.Severity among patients with adverse eventsBoth mild and moderate events4/13 (30.8)—These patients had more than one documented adverse event.Action on upadacitinib among patients with adverse eventsContinued with monitoring7/13 (53.8)—Most non-serious dermatologic, metabolic, and laboratory events were managed without permanent discontinuation.Action on upadacitinib among patients with adverse eventsTemporarily interrupted or held4/13 (30.8)—Temporary interruption was mainly documented for herpesvirus-related infections until clinical control was achieved.Action on upadacitinib among patients with adverse eventsDiscontinued2/13 (15.4)—Discontinuation was documented when infection risk coexisted with laboratory abnormalities or recurrent infection concerns.Documented adverse-event outcome among patients with adverse eventsAll documented events resolved or improved9/13 (69.2)—Events were recorded as resolved, resolved/improved, improved, or stable after management.Documented adverse-event outcome among patients with adverse eventsAt least one event persistent, ongoing, or under follow-up4/13 (30.8)—Persistent or ongoing findings mainly involved lipid/liver-related abnormalities or weight gain. Adverse Events and Laboratory Abnormalities Laboratory monitoring (post baseline) was available for the majority of patients. Laboratory parameters revealed alterations in selected hematological, hepatic, lipid, bilirubin and LDH parameters, with some peak or nadir abnormalities. These findings offer supportive descriptive data to support the use of laboratory monitoring during upadacitinib treatment in real-world clinical practice (Table 6). Table 6Baseline and Peak/Nadir Laboratory Parameters Among Monitored Patients (n=28)ParameterBaseline Median (IQR) [Range]Peak/Nadir Median (IQR) [Range]ANC, x109/L3.1 (2.6–3.5) [1.4–3.9]2.9 (2.4–3.3) [1.2–3.8]ALT, U/L27.0 (23.5–30.0) [18.0–52.0]32.0 (30.3–36.0) [23.0–118.0]AST, U/L23.0 (20.0–26.0) [17.0–44.0]29.0 (25.8–31.3) [21.0–92.0]LDL cholesterol, mg/dL104.0 (95.5–115.3) [84.0–144.0]114.5 (106.8–127.3) [96.0–176.0]Triglycerides, mg/dL112.0 (104.0–128.0) [88.0–180.0]130.0 (122.5–143.0) [98.0–265.0]Bilirubin, mg/dL0.7 (0.6–0.8) [0.5–0.9]0.8 (0.7–0.9) [0.7–1.9]LDH, U/L190.0 (178.0–202.0) [170.0–210.0]212.0 (196.0–222.0) [188.0–386.0]Note: ANC values are presented as baseline and nadir values; other laboratory values are presented as baseline and peak values.Abbreviations: ANC, absolute neutrophil count; ALT, alanine aminotransferase; AST, aspartate aminotransferase; LDH, lactate dehydrogenase; LDL, low-density lipoprotein. Baseline and Peak/Nadir Laboratory Parameters Among Monitored Patients (n=28) Note: ANC values are presented as baseline and nadir values; other laboratory values are presented as baseline and peak values. Abbreviations: ANC, absolute neutrophil count; ALT, alanine aminotransferase; AST, aspartate aminotransferase; LDH, lactate dehydrogenase; LDL, low-density lipoprotein. The five patients who discontinued upadacitinib had diverse reasons for discontinuation, such as safety concerns, infection/laboratory concerns, no meaningful improvement, medication access or loss to follow-up, and unknown after data cleaning. Discontinuation was observed in patients with documented clinical improvement as well as in patients who were not evaluable or had no meaningful improvement, suggesting that treatment discontinuation was not solely due to lack of response (Table 7). Table 7Details of Patients Who Discontinued Upadacitinib (n=5)Dose (mg)TreatmentDuration (Months)Baseline SeverityReason for StoppingStopping CategoryPhysician-Documented Response156.0SevereMild neutropenia 1.7, history of secondary bacterial infectionsSafety/infection or laboratory concernDramatic improvement3012.0SevereNo improvementIneffectiveness/no improvementNo meaningful improvement3012.0ModerateRecurrent infections, herpes zoster, neutropeniaSafety/infection or laboratory concernMajor improvement before discontinuation308.0SevereMedication unavailability and lost follow-upMedication access/lost follow-upNot evaluable: insufficient follow-up157.8SevereUnspecified after removal of Subcorneal pustular dermatosisOther/unclear after cleaningNot evaluable: discontinued early/unclear Details of Patients Who Discontinued Upadacitinib (n=5) Discussion The principal descriptive findings were that qualitative physician-documented improvement was recorded in 27 of the 28 patients with evaluable follow-up and that directly recorded pruritus, sleep-disturbance, and flare-frequency measures decreased during treatment. At the analysis cut-off, 25 of 30 patients remained on upadacitinib, while five had discontinued treatment, including two because of safety, infection, or laboratory concerns. Adverse events were documented in 13 patients. These findings describe clinical observations recorded during upadacitinib treatment but should not be interpreted as establishing that upadacitinib caused the observed improvements. The retrospective, single-centre, non-comparative design, incomplete outcome documentation, unblinded clinical assessment, variable follow-up, and absence of standardized repeated skin-severity measures require cautious interpretation. Guidelines are increasingly acknowledging that while oral Janus kinase inhibitors should not be a first-line treatment for atopic dermatitis (AD), they are an important option for adults or adolescents with moderate-to‑severe disease who have not improved with or cannot tolerate topical therapy, phototherapy, or biologics. The American Academy of Dermatology (AAD) and Joint Task Force (JTF) guidelines only moderately/conditionally recommend upadacitinib and abrocitinib for refractory AD because their high efficacy is tempered by uncertainty regarding long‑term harm and the need for careful monitoring.3 Despite these cautions, real‑world data show that physicians continue to prescribe upadacitinib for patients with a persistent disease burden; in the multi-country AD‑VISE study, 46% of participants were started on 15 mg, the lowest effective dose, whereas 42% received 30 mg because of severe disease or itch. Early responses were notable by two months, 53.6% of patients achieved clear/almost clear skin, and 47% achieved an improvement of ≥90% in their EASI score, responses that increased to 63.0% and 60.7%, respectively, by month six.8 A multicentre real‑world cohort focusing on difficult body sites reported that upadacitinib reduced the mean EASI score from 18.9 at baseline to 2.4 at week 52 and achieved minimal disease activity in 69% of patients.9 Long‑term observational studies support the durability of these outcomes; in a Japanese prospective study of 327 patients, improvements in EASI scores, control and pruritus observed by week 4 were largely maintained over 96 weeks,10 while a Greek single-centre cohort of 36 patients achieved EASI75/90/100 rates of 100/93.8/81.3% at week 48 and maintained clear skin with no pruritus for up to 96 weeks.11 While our results suggest that upadacitinib is predominantly used in patients with significant disease burden or treatment failure and that real‑world outcomes may match those of trials, this descriptive analysis cannot establish comparative efficacy or long‑term safety. Selection bias, the small sample size, and the absence of a control group mandate cautious interpretation. Nevertheless, this study adds to the growing volume of real‑world evidence supporting upadacitinib as a valuable option for moderate‑to‑severe AD when conventional systemic therapy and biologics are inadequate while emphasizing the need for individualized patient selection, baseline screening, and ongoing safety monitoring. The present cohort had a median age of 25 years and a broadly similar proportion of male patients to the pivotal upadacitinib trial populations, although it was younger than the pooled Measure Up population. The Measure Up trials used standardized eligibility criteria, including EASI, validated Investigator Global Assessment for Atopic Dermatitis, affected body surface area, and Worst Pruritus Numerical Rating Scale thresholds. In contrast, baseline disease severity in the present cohort was based on routine physician classification without documented baseline EASI, vIGA-AD, or body surface area measurements. Demographic characteristics can therefore be described in relation to pivotal trial populations, but direct comparison of baseline disease severity and standardized treatment response is not possible. In this retrospective cohort, most patients (27/28) demonstrated some degree of physician‑documented improvement. The median time to first response was four weeks, with seven patients improving within two weeks and 16 patients showing early response by 3–4 weeks. These observations parallel the rapid onset of action reported in Phase 3 trials. An integrated analysis of the double‑blind portions of Measure Up 1 and 2 showed that a clinically meaningful itch reduction (≥4‑point improvement on worst‑pruritus NRS) occurred as early as week 1 in patients treated with upadacitinib 15 mg (11.2%) or 30 mg (17.7%), whereas only 0.5% of placebo recipients achieved this benchmark. Subsequent improvements in itch, skin pain, sleep, activities, and emotional state were evident within 1–2 weeks and were maintained through week 16. A combined analysis of Measure Up and the head‑to‑head Heads Up trial further demonstrated that upadacitinib‑treated patients achieved WP‑NRS 0/1 by day 2, EASI 90 responses by week 1, and composite EASI 90 + WP‑NRS 0/1 by week 2, with the benefits persisting through weeks 16–24.12 Real-world studies also suggest early improvements: the TREATgermany registry reported improvements in EASI 75/90 scores in 63%/40% of adults by three months, and a single-centre study documented marked reductions in EASI and itch NRS scores by week 8.13 The rapid physician‑documented responses seen in our cohort thus align with the quick symptom relief observed in clinical trials and real‑world registries. Rapid amelioration of itch and skin inflammation is clinically meaningful for patients with a high symptom burden, as it may reduce sleep disturbance and improve quality of life. Nevertheless, our findings should be interpreted cautiously because they rely on physician notes rather than validated outcome measures, the cohort is small, and the retrospective non‑comparative design cannot account for placebo effects or natural disease fluctuation. An important methodological distinction is that the overall response categories—dramatic, major, moderate, partial, or no meaningful improvement—were derived by the research team from routine narrative clinical documentation and lacked predefined quantitative thresholds. These categories should therefore not be equated with standardized outcomes such as EASI-75, EASI-90, or vIGA-AD 0/1. In contrast, 0–10 pruritus numerical rating scale scores and 0–10 sleep-disturbance ratings were directly recorded in the medical records and permitted quantitative within-patient assessment. Nevertheless, a uniform follow-up assessment time was unavailable, and the validation status of the sleep-disturbance rating could not be established. The symptom findings therefore remain descriptive and cannot be directly compared with protocol-defined clinical-trial endpoints. In our retrospective cohort, symptom improvement was recorded for itch, sleep, and flare burden: Among 28 patients with paired observations, the median pruritus NRS decreased from 8.0 to 2.0, the median sleep-disturbance rating decreased from 8.0 to 2.0, and monthly flare frequency decreased from 5.0 to 1.0. A reduction of at least four points in pruritus NRS was recorded in 25 patients. These symptoms are central drivers of quality-of-life impairment in moderate to severe atopic dermatitis (AD). Randomized trials demonstrate that upadacitinib improves itch and sleep rapidly. In pooled phase 3 trials, ≥ 4-point reductions in worst-pruritus NRS were observed as early as week 1 and increased through weeks 4–6, while improvements across patient reported outcomes, including skin pain, sleep (ADerm IS sleep), daily activities and mental health, occurred within 1–2 weeks and were sustained through week 16.12 At week 1, a clinically meaningful improvement in sleep was reported by 16.9% and 21.7% of upadacitinib 15 mg and 30 mg recipients, respectively, versus 1.3% with placebo,12 while a 52-week extension study showed meaningful sleep improvements exceeding 15–20% at week 1 and 50–65% at week 4, which were maintained through week 52.14 Real‑world data mirror these findings: in the UP‑TAINED study, mean sleep‑loss scores fell from 5.8 to 1.7 within three months, and 77% of patients achieved ≥50% sleep improvement at one month, while 47% achieved itch scores of 0/1 and 68.7% had a ≥4‑point reduction in pruritus.15 Our physician‑documented reductions in itch, sleep disturbances, and flares align with these therapeutic goals. However, these improvements were recorded retrospectively and lack validated symptom instruments; therefore, they should be interpreted cautiously and not as definitive evidence of efficacy. Several alternative explanations for the observed changes should be considered. Response-related analyses were limited to patients with evaluable follow-up, and routine clinical documentation may have preferentially captured more evident improvement, introducing outcome-availability and documentation bias. The absence of a comparator group means that natural disease fluctuation, regression to the mean, changes in adherence or supportive care, and temporal, seasonal, or environmental influences cannot be separated from changes associated with upadacitinib exposure. In addition, concomitant topical or systemic treatments were not consistently captured and may have contributed to improvements in pruritus, sleep disturbance, flare frequency, or overall clinical status. The observed reductions should therefore be described as changes recorded during treatment rather than effects attributable solely to upadacitinib. This cohort offers real-world insight into treatment continuation with upadacitinib in routine practice. At the analysis cut-off, treatment was ongoing in five-sixths of patients, with a median treatment duration of 12 months, and there were only five discontinuations (median duration of eight months). Published real-world studies show that persistence with upadacitinib is generally high, but discontinuations do occur. In an Italian multicentre cohort (325 adults), 1-year and 1.5-year drug survival rates for upadacitinib were 91.5% and 80.2%; only 7.7% of patients discontinued the medication, mainly because of adverse events (4.9%), including infections, laboratory abnormalities, thromboembolism, and malignancy, while inefficacy accounted for 0.6%.16 By contrast, an international registry of 549 patients treated with biologics or JAK inhibitors reported that upadacitinib was discontinued in 44% of cases; the median treatment duration when continued was seven months, and discontinuation was most often due to ineffectiveness (24.3%) or adverse events (25.2%), such as upper respiratory infections, acne, or headache.17 Long‑term extension data indicate that upadacitinib maintains its efficacy up to 160 weeks, with low rates of serious adverse events and discontinuations.18 These findings suggest that continued therapy may be necessary to sustain disease control. Our retrospective analysis mirrors these observations: discontinuations occurred for safety concerns, infection or laboratory abnormalities, perceived lack of benefit, and access issues. Treatment persistence in our cohort likely reflects a combination of clinical response, tolerability, and patient motivation rather than definitive evidence of sustained efficacy. Real-world persistence is influenced by monitoring, comorbidities, adverse events, adherence, and drug access;19 therefore, the results of this study should be interpreted cautiously and cannot substitute for controlled efficacy or safety outcomes. Nevertheless, the ongoing treatment of most patients despite prior systemic therapy exposure underscores the clinical utility of upadacitinib in a difficult-to-treat population. Larger prospective studies are needed to identify predictors of persistence and disentangle the relative contributions of efficacy, safety, and access. Because our analysis was retrospective, non-comparative, and subject to variable follow-up durations, the treatment persistence observed should be considered exploratory and not proof of superiority over other advanced systemic agents. In this single-centre cohort of 30 patients, whose median age was 25 years and median treatment duration was 12 months, six patients experienced infection-related adverse events, including herpes zoster and herpes simplex or eczema herpeticum. Neutropenia and lipid- and liver-related abnormalities were also documented. This pattern is consistent with established adverse events in the AD upadacitinib programme. In the six-year integrated trial analysis, exposure-adjusted event rates for the 15- and 30-mg doses, respectively, were 2.2 and 2.6 per 100 patient-years for serious infection, 3.2 and 5.2 for herpes zoster, 1.5 and 1.9 for eczema herpeticum, and 3.4 and 5.6 for oral herpes.7 Hepatic disorders and neutropenia were also numerically more frequent with 30 mg than with 15 mg, as were several other adverse events, including acne. A meta-analysis of 50 real-world AD studies similarly identified acne, creatine phosphokinase elevation, lipid elevation, and herpesvirus infections among the most frequently reported adverse events.20 However, its pooled proportions were heterogeneous, mostly derived from single-arm studies, and should not be directly compared with exposure-adjusted clinical-trial incidence rates. A propensity-matched AD cohort also found higher risks of herpes simplex, herpes zoster, skin infection, neutropenia, and hyperlipidaemia with oral JAK inhibitors than with dupilumab over a maximum follow-up of two years.21 Collectively, these findings support regular infection surveillance and laboratory monitoring rather than suggesting that the events observed in the present cohort were unexpected. One patient had clinically documented baseline ethnic neutropenia. Previous Saudi studies have reported isolated benign neutropenia among otherwise healthy individuals, with variation across geographic regions. However, genetic ancestry, Duffy-null status, and population-specific absolute neutrophil count reference intervals were not assessed in the present cohort. Consequently, low or changing neutrophil counts could not always be confidently distinguished as pre-existing population-associated variation, a treatment-related change, or a combination of both. This uncertainty also limits the generalisability of the neutropenia findings to populations with different baseline neutrophil-count distributions.22 No MACE, venous thromboembolism, malignancy, or death was documented in the present cohort. This finding should be interpreted as an absence of recorded events, not as evidence that these risks were absent. The study included only 30 patients, follow-up was variable, and total accrued patient-years were not reported; it therefore had limited capacity to detect uncommon or delayed outcomes. In the six-year AD trial analysis, MACE occurred at approximately 0.2 and <0.1 events per 100 patient-years with upadacitinib 15 and 30 mg, respectively, while venous thromboembolism occurred at approximately 0.1 event per 100 patient-years with both doses.7 A subsequent correction clarified that 11 venous thromboembolic events occurred, six were serious, and none were fatal.23 Observational findings are not entirely uniform. One propensity-matched study did not detect significantly increased MACE, venous thromboembolism, malignancy, or mortality with oral JAK inhibitors compared with dupilumab, but follow-up was limited and the analysis combined multiple JAK inhibitors.21 Conversely, another global cohort reported increased pulmonary embolism and deep-vein thrombosis risks relative to dupilumab and an increased deep-vein thrombosis risk relative to methotrexate, although the absolute risk differences were small and residual confounding cannot be excluded.24 These findings reinforce the need to distinguish low absolute event frequencies from potentially important relative risks in susceptible patients. The young age distribution of the present cohort further limits extrapolation to older or cardiovascular-risk populations. Only 116 participants aged ≥65 years were included across both dose groups in the six-year AD trial analysis, representing approximately 338 patient-years of exposure. Serious adverse events, treatment discontinuations, serious infections, and several cardiovascular or malignancy outcomes were numerically more frequent in this subgroup, but estimates were imprecise because of the small sample and wide confidence intervals.7 Separately, a Danish registry study found that approximately half of adults with AD had at least one study-defined factor potentially affecting JAK-inhibitor selection, with multiple risk factors becoming more common with increasing age.25 This proportion should not be interpreted as the percentage of patients with a formal contraindication. In the present study, the median age of 25 years may partly explain the absence of rare age-associated outcomes; however, low risk cannot be assumed unless smoking history, cardiovascular disease, hypertension, diabetes, obesity, dyslipidaemia, prior malignancy, previous thromboembolism, and recurrent or serious infection were systematically assessed and reported. The present safety observations should therefore be interpreted within this small, predominantly young treated cohort while remaining descriptive and hypothesis-generating rather than evidence that regulatory warnings are inapplicable. There are several strengths to this study. First, it offers real-world data on the use of upadacitinib in patients with moderate-to-severe atopic dermatitis in routine clinical practice. This cohort comprised patients with known prior exposure to treatment, including topical therapy, oral corticosteroids, phototherapy, cyclosporine, dupilumab, and other systemic or biologic therapies, reflecting the clinical complexity that is often seen in practice. Second, the analysis included multiple clinically relevant domains such as treatment exposure, treatment duration, physician documented response, timing of response, symptom improvement, discontinuation, adverse events, and laboratory monitoring. Third, safety surveillance was relatively well documented, with post-baseline laboratory monitoring available for most patients, allowing investigation of abnormalities in parameters like absolute neutrophil count, liver enzymes, lipid parameters, bilirubin, and lactate dehydrogenase. The provision of detailed reasons for discontinuation further enhances the clinical utility of the results, as treatment continuation was not only related to response, but also to safety issues, infections, access to medication and reliability of follow-up. However, several limitations should be considered. This was a retrospective, descriptive study involving only 30 patients from a single centre. Individual observations therefore materially affected the percentages, uncommon or delayed outcomes could not be assessed reliably, and stable dose-, age-, comorbidity-, or previous-treatment subgroup analyses were not possible. The single-centre setting may reflect local referral, prescribing, monitoring, follow-up, and medication-access practices that are not representative of other Saudi or international settings. Treatment allocation was non-random, and the determinants of prescribing upadacitinib were not systematically captured; prescription bias and confounding by indication therefore cannot be excluded. Patients selected for upadacitinib may have differed from other patients with atopic dermatitis in disease burden, previous treatment history, comorbidities, access to treatment, or clinician-assessed suitability. Response analysis was also susceptible to outcome-availability, documentation, and detection bias. Two patients lacked sufficiently clear follow-up documentation for response evaluation, routine clinical notes may have preferentially captured more evident changes, and the treating clinicians were aware that patients were receiving upadacitinib. Whether the same clinician performed all follow-up assessments could not be confirmed, and no independent blinded adjudication was undertaken. Repeated standardized skin-severity and quality-of-life measures, including EASI, vIGA-AD, SCORAD, POEM, and DLQI, were unavailable. Although directly recorded pruritus NRS and sleep-disturbance ratings permitted paired descriptive assessment, follow-up timing was not standardized and the validation status of the sleep rating could not be established. Furthermore, concomitant topical or systemic treatments, adherence changes, and seasonal or environmental exposures were not consistently captured. Natural disease fluctuation, regression to the mean, and concurrent interventions therefore cannot be separated from changes occurring during upadacitinib treatment. The retrospective records may also have incompletely captured adverse events and previous treatment exposure. BMI and smoking status were unavailable, preventing full characterization of cardiovascular and thromboembolic risk. One patient had documented baseline ethnic neutropenia, but genetic ancestry, Duffy-null status, and population-specific neutrophil reference intervals were not assessed; consequently, baseline population-associated variation could not be separated confidently from treatment-related neutrophil changes. The apparently low proportion with documented previous topical therapy likely reflects incomplete treatment histories and should not be interpreted as indicating that the remaining patients had never received topical treatment. Finally, the small numbers previously exposed to dupilumab, baricitinib, or another biologic prevented meaningful evaluation of the influence of previous advanced therapy. These limitations should be addressed in future studies with larger numbers of patients in prospective multicentre studies with standardized baseline and follow-up assessments. These studies should incorporate validated clinician and patient-reported outcome measures such as EASI, SCORAD, POEM, DLQI, itch NRS, and sleep-related instruments to allow for more comprehensive assessment of clinical response and symptom burden. There is also a need for longer follow-up to evaluate treatment persistence, durability of response, discontinuation rates, and delayed safety outcomes. Laboratory abnormalities should be studied in the future with predefined grading criteria and infection-related events such as herpes zoster, herpes simplex, and eczema herpeticum should be systematically documented. Comparative real-world studies of upadacitinib versus biologics, other oral JAK inhibitors, or traditional systemic therapies would also be useful in understanding the relative position of upadacitinib in treatment pathways. Lastly, predictors of response, discontinuation, adverse events, and treatment persistence, especially in patients with previous biologic use, comorbidities, and baseline laboratory abnormalities, should be studied. Conclusion In this small single-centre cohort, physician-documented improvement and reductions in directly recorded pruritus, sleep-disturbance, and flare-frequency measures were observed during upadacitinib treatment. Because the study was retrospective, non-comparative, and susceptible to prescription, documentation, observer, and concomitant-treatment bias, these findings cannot establish that upadacitinib caused the observed improvements. However, five patients discontinued treatment, including two because of safety, infection, or laboratory concerns, and the absence of standardized skin-severity assessments, independent blinded evaluation, uniform follow-up timing, and complete cardiovascular-risk information limits interpretation. These findings support individualized patient selection, baseline risk assessment, and structured clinical and laboratory monitoring. Larger prospective multicentre studies using standardized clinician-reported and patient-reported outcomes are required. Acknowledgments I am grateful for Dr. Malak Fuad Alshaebi, King Fahd University Hospital, for her support and cooperation during the study. Disclosure The author declares no conflicts of interest. References