Real-world clinical outcomes, prognostic factors, and safety profile of upadacitinib in patients with inflammatory bowel disease

Abstract

BACKGROUND

Upadacitinib has been approved for the treatment of patients with moderate-to-severe Crohn’s disease (CD) and ulcerative colitis (UC). Observational studies on the use of upadacitinib in inflammatory bowel disease (IBD) outside regulatory trials remain limited.

AIM

To evaluate real-world effectiveness and adverse events (AEs) profile of upadacitinib in patients with IBD.

METHODS

This is an observational cohort study in patients with refractory IBD [CD: 43; UC: 42; indefinite colitis: 1; pouchitis: 1; mean age: 40.1 years (SD = 12.9); median disease duration: 6 years (interquartile range: 2.5-13); extra-intestinal manifestations: 44 (50.6%)], who received treatment with upadacitinib. Our primary endpoint was treatment persistence. Secondary endpoints included clinical remission, corticosteroid- and intensification-free treatment persistence and endoscopic remission rates. AEs were also reported.

RESULTS

Eighty-six patients were exposed to ≥ 1 advanced therapy and 40.2% received concomitant corticosteroids at baseline. Seventy patients (80.5%) remained on upadacitinib (median duration: 66 weeks, interquartile range: 30-88) by the end of follow-up. Treatment persistence was higher in UC (88.1% vs 72.1% in CD, P = 0.046). Clinical remission at one-year was achieved in 63.3%, corticosteroid-free remission in 51.7%, and intensification-free remission in 58.3%. In patients with UC, smoking (P = 0.014) and baseline C-reactive protein (P = 0.029) were associated with earlier treatment discontinuation, whereas, male sex (P = 0.037) and extensive disease (P = 0.045) with earlier discontinuation or need for intensification. Endoscopic remission was achieved in 41.2% (7/17) of CD and 65.2% (15/23) of UC patients. In multivariate regression, endoscopic remission was associated with female sex (P = 0.039) and UC (P = 0.041). AEs were reported in 48 patients, primarily infections (24.1%) and acne (21.8%).

CONCLUSION

In this real-world cohort of patients with IBD who were refractory to multiple prior advanced therapies, upadacitinib demonstrated high treatment persistence and satisfactory clinical remission rates.

Key Words: Upadacitinib; Inflammatory bowel disease; Ulcerative colitis; Crohn’s disease; Treatment; Drug sustainability; Efficacy

Core Tip: Our real-world study demonstrates that upadacitinib is an effective and well-tolerated treatment for patients with advanced therapy-refractory inflammatory bowel disease. Over a median follow-up of 66 weeks, upadacitinib showed high treatment persistence and clinical remission rates, particularly in patients with ulcerative colitis. Moreover, our study identifies baseline predictors for treatment discontinuation or intensification of upadacitinib in patients with ulcerative colitis, as well as predictors of endoscopic remission in both ulcerative colitis and Crohn’s disease, thus, providing insights into personalized management of patients.

INTRODUCTION

Inflammatory bowel disease (IBD), which consist of Crohn’s disease (CD), ulcerative colitis (UC) and indeterminate colitis (IC), are immune-mediated disorders that primarily affect the gastrointestinal tract. They are characterized by chronic inflammation of the large and small intestines, which, if left uncontrolled, can lead to progressive bowel damage[1,2].

Over the past three decades, the therapeutic landscape of IBD has evolved substantially with the introduction of several targeted therapies, which include biologic agents (anti-tumor necrosis factor α, anti-α4β7 integrin, anti-interleukin (IL)-12/23, anti- IL-23) and small molecules [Janus kinase (JAK) inhibitors and S1P modulators] and are collectively referred to as advanced therapy[3]. Despite this considerable progress, however, primary and secondary treatment failure remain high for most available drugs, leading to persistent disease activity, corticosteroid dependence, hospitalization, and surgery in a sizeable proportion of patients[4]. Furthermore, treatment targets in IBD have evolved towards the accomplishment of objective markers of the elimination of inflammation, in particular endoscopic, histologic and transmural healing, which all indicate optimal disease control, according to the Selecting Therapeutic Targets in IBD (STRIDE) initiative[5]. Those new priorities, alongside the increasingly recognized therapeutic complexity and high disease burden for patients with IBD, necessitate the need for treatments that provide rapid symptom control and durable remission, combined with favorable safety profiles.

JAK inhibitors have recently emerged as an important therapeutic class in IBD. They are orally administered small molecules with a rapid onset of action. Upadacitinib has increased selectivity for JAK1 and is currently the only member of this class of drugs that has been approved for the treatment of both UC and CD[6]. In particular, in moderate-to-severe UC, the phase III U-ACHIEVE and U-ACCOMPLISH trials reported significantly improved clinical, biochemical, endoscopic, and histologic outcomes in patients treated with upadacitinib, as compared to placebo[7]. Similarly, phase III U-EXCEL, U-EXCEED and U-ENDURE studies, and, more recently, the U-ENDURE extension demonstrated the efficacy of upadacitinib in patients with moderate-to-severe CD, with previous biologic failure[8,9]. Although regulatory trials have established the clinical applicability of upadacitinib in patients with IBD, real-world studies will be of paramount importance to further confirm the efficacy and, also, validate the safety of longitudinal administration in patients with CD or UC.

The aim of the present study was to evaluate the real-world effectiveness and safety of upadacitinib in patients with IBD, who were treated in two Greek tertiary centers. We report, herein, that upadacitinib proved to be a reliable treatment for both CD and UC in a predominantly biologic-experienced patient population and that its administration was associated with high rates of important and up-to-date treatment targets, such as clinical and endoscopic remission, and corticosteroid-free therapy. We were also able to identify predictors of treatment discontinuation and endoscopic remission, which may help to guide future therapeutic decisions.

MATERIALS AND METHODS

Study design

This is an observational cohort study conducted at two tertiary centers in Greece (“Sotiria” Hospital and General Hospital of Athens “Alexandra”). We included adult patients diagnosed with active moderate-to-severe IBD who initiated upadacitinib from October 2022 to March 2026. All patients with a confirmed diagnosis of IBD who initiated upadacitinib in routine clinical practice during the aforementioned time period were eligible for enrollment. Exclusion criteria included the concurrent administration of another advanced therapy (such as an anti-IL-23) and loss to follow-up immediately after baseline. The initiation of upadacitinib, as well as subsequent treatment dosing adjustments or discontinuations, was made at the discretion of treating physicians based on clinical disease activity. Patients with CD were treated with 45 mg daily for the 12-week induction period, and patients with UC with 45 mg daily for the 8-week induction period, and then commenced on 30 mg or 15 mg daily.

Study endpoints

Our primary endpoint was treatment persistence, which was defined as remaining on upadacitinib at one year from initiation. Secondary endpoints were corticosteroid-free treatment persistence, intensification-free treatment persistence, and for those with a recorded visit at 52 week, one-year clinical remission, one-year corticosteroid-free clinical remission, and endoscopic remission rates.

For the definition of clinical remission, we used the patient-reported outcomes (PROs) of “stool frequency” and “abdominal pain” for CD; “stool frequency” and “rectal bleeding” for UC according to STRIDE-II[5]. For CD, clinical remission was defined as PRO “abdominal pain” ≤ 1 and PRO “stool frequency” ≤ 3, while for UC, clinical remission was defined as a sum of PROs = 0. Corticosteroid exposure was defined as any initiation of systemic and/or topical corticosteroids after the initiation of upadacitinib. For those who received concomitant corticosteroids at initiation, the need for dose re-escalation during the tapering of corticosteroids was also considered as exposure to corticosteroids. As treatment intensification was considered the need for dose adjustment to 45 mg daily beyond the induction period, either in the form of prolonged induction, or as re-induction while on maintenance, or the need for the optimization of maintenance dosage of upadacitinib from 15 mg to 30 mg daily. Endoscopic remission was defined as a Simple Endoscopic Score for CD of ≤ 3 with absence of aphthous ulcers in patients with CD with no prior history of disease-related abdominal surgery, as modified Rutgeert’s score ≤ i2a for patients with CD and prior history of disease-related abdominal surgery, and endoscopic Mayo sub-score of 0 for patients with UC.

Ethics statement

The study protocol was approved by the Institutional Review Board of “Sotiria” Hospital, approval Protocol No. 11493, date of approval: April 29, 2026. All enrolled patients had previously provided a general written informed consent during their routine clinical care, authorizing the anonymous recording and retrospective analysis of their medical records for research purposes. Because of the retrospective design of this study, we did not require a new, study-specific informed consent. The study was conducted in accordance with the principles of the Declaration of Helsinki.

Collected data

Age, gender, smoking habits, and disease-related data were collected before upadacitinib initiation. Namely, we recorded the disease type, disease duration, disease extension according to Montreal classification[10], history of surgeries related to CD, presence of extra-intestinal manifestations as well as the prior and concomitant IBD treatments.

For the evaluation of clinical activity in patients with CD, we used the PROs of “abdominal pain” and “stool frequency”[11], and, in patients with UC, we used the PROs of “stool frequency”, “rectal bleeding” and “urgency”[12].

Biochemical activity was evaluated with C-reactive protein (mg/L). For the assessment of endoscopic activity, the Simple Endoscopic Score for CD was recorded in patients with CD[13], and the modified Rutgeert’s score for post-operative CD patients[14], and the endoscopic Mayo sub-score[15] and the UC Endoscopic Index of Severity in patients with UC[16].

Statistical analysis

Statistical analysis was performed with the statistical package SPSS 23 (IBM, Armonk, NY, United States). For categorical variables total count and percentages are presented. For continuous variables that are normally distributed “Mean value” and “Standard Deviation”, while for those not normally distributed “Median” and “Range” are presented. The non-parametric χ² and Fisher’s exact test were used for the comparison of categorical outcomes. Univariate Cox regression and logistic models were used to identify factors related to primary and secondary outcomes. For multivariable regression analyses, covariates were selected based on a combination of statistical significance in univariate screening (e.g., P < 0.10) and established clinical relevance a priori. Demographic factors and disease subtype (UC vs CD) were systematically included in the multivariable models to adjust for clinically significant confounding, regardless of their initial univariate P values. A P value = 0.05 was used as threshold of statistical significance.

RESULTS

Study population

The clinical and demographic characteristics of patients are shown in Table 1. In total we included 87 patients [CD, 43 (49.4%); UC, 42 (48.3%); one patient with IC and one with pouchitis]. Mean age was 40.1 years (SD = 12.9), and 44 (50.6%) were male. Median disease duration was 6 years [interquartile range (IQR): 2.5-13]. Forty-four patients (50.6%) had a history of extra-intestinal manifestations, the most common being articular (34/39.1%). Twenty-seven patients (31%) were active smokers. Regarding previous treatments, 86 patients (98.9%) had a history of exposure to at least 1 advanced therapy, with a median number of 2 (IQR: 1-3) [79 (90.8%) at least one anti-tumor necrosis factor agent, 43 (49.4%) ustekinumab, 21 (24.1%) vedolizumab, and 8 (9.2%) tofacitinib]. Concomitant corticosteroids at upadacitinib commencement were administered to 35 (40.2%) patients (Table 1). The majority of patients in our cohort who maintained on upadacitinib after the induction period commenced with 30 mg daily (92.5% in UC, 92.1% in CD).

Table 1 Demographics and disease-related characteristics of 87 patients initiating upadacitinib, n (%).

Characteristics	n = 87
Age in years, mean (SD)	40.1 (12.9)
Male sex	44 (50.6)
Disease duration in years, median (IQR)	6 (2.5-13)
IBD type
Crohn’s disease	43 (49.4)
Ulcerative colitis	42 (48.3)
Indeterminate colitis	1 (1.1)
Pouchitis	1 (1.1)
Montreal classification
UC - E (1/2/3)	4 (9.5)/17 (40.5)/21 (50.0)
CD - A (1/2/3)	3 (7.0)/32 (74.4)/8 (18.6)
CD - L (1/2/3)	16 (37.2)/6 (14.0)/21 (48.8)
CD- B (1/2/3)	23 (53.5)/18 (41.9)/2 (4.6)
CD -L4	4 (9.3)
CD - p	12 (27.9)
History of IBD-related surgery in CD	17 (39.5)
Extraintestinal manifestations	44 (50.6)
Articular manifestations	34 (39.1)
Dermatologic manifestations	16 (18.4)
Occular manifestations	2 (2.3)
Hepatobiliary manifestations	0 (0.0)
Other	9 (10.3)
Prior AT exposure	86 (98.9)
Number of prior AT-0/1/2/3/4	1 (1.1)/23 (26.4)/35 (40.2)/26 (29.9)/2 (2.3)
Median number of prior advanced therapies, (IQR)	2 (1-3)
Prior anti-TNF exposure	79 (90.8)
Prior ustekinumab exposure	43 (49.4)
Prior vedolizumab exposure	21 (24.1)
Prior tofacitinib exposure	8 (9.2)
Concomitant corticosteroids at initiation	35 (40.2)

CD: Crohn’s disease; IBD: Inflammatory bowel disease; IQR: Interquartile range; TNF: Tumor necrosis factor; UC: Ulcerative colitis; AT: Advanced therapy.

Efficacy outcomes were evaluated exclusively in patients with confirmed diagnoses of UC or CD. Patients with indefinite colitis or pouchitis (n = 1 each), were included only in the overall safety population, to provide comprehensive real-world safety reporting for all upadacitinib-exposed individuals in our cohort. A detailed flow of patients through the study is available in Supplementary Figure 1.

Clinical efficacy data

At the end of follow-up, 70 patients (80.5%) were still on upadacitinib, with a median treatment duration of 66 weeks (IQR: 30-88). Patients with UC had significantly higher treatment persistence rates (88.1%) than patients with CD (72.1%) (Log-rank P = 0.046, Figure 1). Corticosteroid-free upadacitinib persistence was observed in 51 (58.6%) patients. The rates of corticosteroid-free treatment persistence and intensification-free treatment persistence did not differ between patients with UC and CD (Supplementary Figure 2). Prolonged induction was implemented in 7 patients (all with UC), and re-induction in 5 (2 with UC and 3 with CD).

Figure 1 Kaplan-Meier curve of upadacitinib persistence. The blue line represents patients with ulcerative colitis, and the green line represents patients with Crohn’s disease. Patients with ulcerative colitis achieved significantly higher persistence rates compared to those with Crohn’s disease (P = 0.046). Numbers beside the curves indicate the proportion of persistent patients at the end of follow-up for each group. UC: Ulcerative colitis; CD: Crohn’s disease; CI: Confidence interval.

At one-year on treatment, among 60 patients with sufficient follow-up period, clinical remission was recorded in 38 (63.3%) and one-year corticosteroid-free clinical remission in 31 (51.7%) (patients with pouchitis and IC were excluded from this analysis). Rates of clinical remission and one-year corticosteroid-free clinical remission were higher in patients with UC (74.2% and 54.8%, respectively) compared to patients with CD (51.7% and 48.3%, respectively), although not to a statistically significant level (Figure 2).

Figure 2 One-year efficacy outcomes of upadacitinib in patients with ulcerative colitis and Crohn’s disease. Dark blue bars represent the proportion of patients achieving each outcome, with exact patient numbers indicated at the base of each bar. Patients with ulcerative colitis achieved numerically higher rates compared to patients with Crohn’s disease. A: One-year clinical remission rates; B: One-year corticosteroid-free clinical remission rates; C: Endoscopic remission rates. UC: Ulcerative colitis; CD: Crohn’s disease.

Endoscopic efficacy data

Follow-up endoscopic evaluation was available for 40 patients. In all assessed patients, follow-up endoscopy was performed at least 6 months after initiation of upadacitinib treatment. The primary reasons for not undergoing post-treatment endoscopy included patient refusal due to clinical asymptomatic remission, and ongoing active disease, leading to early treatment discontinuation before scheduled endoscopic re-assessment. Patients who did not undergo a follow-up endoscopic evaluation differ only regarding their Montreal classification, having higher prevalence of ileal disease (Supplementary Table 1).

Endoscopic remission was recorded in 15/23 (65.2%) of patients with UC and 7/17 (41.2%) of patients with CD (P = 0.131) (Figure 2). Interestingly, among patients with CD, endoscopic remission was recorded in 100% (4/4) of patients with colonic CD compared to 23.1% (3/13) of patients with involvement of the terminal ileum (ileal, or ileo-colonic location) (Fischer’s exact test P = 0.015).

Safety data

Eighty-seven adverse events (AEs) were recorded in 48 (55.2%) patients. Infectious complications were the most frequent AE in our cohort, affecting 21 patients (24.1%) who experienced at least one such event during the follow-up period. Other commonly observed AEs were acne (n = 19, 21.8%), alterations in lipid profile parameters (n = 15, 17.2%), and increased serum creatine phosphokinase levels (n = 12, 13.8%). Two patients experienced Herpes-Zoster. In total, 5 patients (5.7%) discontinued upadacitinib because of an AE, specifically due to transaminase elevation (n = 3), hidradenitis suppurativa flare (n = 1), and erythema multiforme (n = 1). Hospitalization due to an AE was required in four patients (4.6%). Importantly, no major adverse cardiovascular events (MACE), thromboembolic events or deaths have been reported in our cohort at the time of analysis. Figure 3 depicts the temporal distribution of AEs under treatment with upadacitinib. Interestingly, most AEs occurred during the first 20 weeks following drug initiation, while the incidence of new-onset events decreased substantially in later time-points for the majority of categories. Alterations in liver enzymes and lipidemic profile, as well as acne, were predominantly observed in the early phases of treatment, while infections showed scattered distribution, that persisted throughout the 140-week follow-up period. A summary of AEs reported in our cohort is provided in Supplementary Table 2.

Figure 3 Time distribution of adverse events during treatment with upadacitinib. The horizontal axis represents time in weeks after upadacitinib initiation. The vertical axis categorizes the recorded adverse events (AEs). Each data point denotes an individual event in each AE category, illustrating the onset and frequency of specific side effects recorded throughout the follow-up period. Symbols represent different AE categories. Blue circles: Infections; Blue triangles: Acne; Blue cross marks: Lipid profile alterations; Pink asterisks: Creatine phosphokinase levels elevation; Purple diamonds: Transaminase elevation; Orange plus signs: Other AEs. Other AEs: Hidradenitis suppurativa flare (one patient at week 7); erythema multiforme (one patient at week 13); hair loss - two patients at weeks 14 and 15); skin rash (one patient at week 39); elevated creatinine (one patient at week 44); migraine (one patient at week 56); leukopenia (one patient at week 79).

Predictors of outcomes

No significant predictors of treatment outcomes were identified among patients with CD. In contrast, among patients with UC, univariate Cox regression showed that earlier discontinuation of upadacitinib significantly correlated with active smoking [hazard ratio (HR) = 9.40, 95% confidence interval (CI): 1.51-58.6; P = 0.016] and baseline C-reactive protein levels (HR = 1.08, 95%CI: 1.01-1.16; P = 0.023). Both factors remained significant in multivariable Cox regression analysis [active smoking, HR = 11.37, 95%CI: 1.64-78.95; P = 0.014; baseline C-reactive protein (CRP) levels, HR = 1.09, 95%CI: 1.01-1.18; P = 0.029] as independent predictors of earlier treatment discontinuation. Similarly, in univariate Cox regression earlier treatment discontinuation or intensification of upadacitinib was significantly associated with male sex (HR = 4.76, 95%CI: 1.06-20; P = 0.041) and showed a trend for association with extensive disease (Montreal classification E3) (HR = 2.95, 95%CI: 0.92-9.42; P = 0.068). Male sex (HR = 5, 95%CI: 1.10-20.00; P = 0.037) and extensive disease (Montreal classification E3) (HR = 3.30, 95%CI: 1.03-10.58; P = 0.045) were independently associated with earlier treatment discontinuation or the need for treatment intensification in multivariate Cox regression (Table 2).

Table 2 Cox regression of factors associated with upadacitinib discontinuation and discontinuation or intensification in patients with ulcerative colitis (n = 42).

n = 42	Univariate			Multivariate
Discontinuation (17 events)	HR	95%CI	P value	HR	95%CI	P value
Smoking	9.40	1.51-58.6	0.016	11.37	1.64-78.95	0.014
CRP (mg/L)	1.08	1.01-1.16	0.023	1.09	1.01-1.18	0.029
Discontinuation or intensification (27 events)	HR	95%CI	P value	HR	95%CI	P value
Female	0.21	0.05-0.94	0.041	0.20	0.05-0.91	0.037
Extensive colitis (E3)	2.95	0.92-9.42	0.068	3.30	1.03-10.58	0.045

CI: Confidence interval; CRP: C-reactive protein; HR: Hazard ratio.

Although no factor was significantly associated with achievement of endoscopic remission in univariate logistic regression, a tendency was identified for female sex [odds ratio (OR): 2.75, 95%CI: 0.76-9.95; P = 0.123] and UC (OR: 2.70, 95%CI: 0.74-10; P = 0.135). Since both of these factors were previously associated with superior upadacitinib response (Table 2 and Figure 1) we analyzed them in a multivariate model. In multivariate logistic regression, female sex (OR: 5.92, 95%CI: 1.10-32.0; P = 0.039) and a diagnosis of UC (OR: 5.88, 95%CI: 1.08-33.33; P = 0.041) were independently associated with a higher likelihood of achieving endoscopic remission (Table 3).

Table 3 Logistic regression of factors associated with endoscopic remission (n = 23).

Characteristics	Univariate			Multivariate
	OR	95%CI	P value	OR	95%CI	P value
Female	2.75	0.76-9.95	0.123	5.92	1.10-32.0	0.039
Crohn’s disease	0.37	0.10-1.36	0.135	0.17	0.03-0.93	0.041

CI: Confidence interval; OR: Odds ratio.

DISCUSSION

Herein, we present real-world evidence on the efficacy of upadacitinib in patients with IBD who had previously failed multiple advanced therapies. We report that, in our cohort, upadacitinib demonstrated high rates of treatment persistence, favorable effectiveness in clinical and endoscopic targets, as well as an acceptable safety profile.

In our cohort, more than 80% of patients remained on treatment with upadacitinib after a median follow-up of 66 weeks. We selected treatment persistence as the primary outcome as it encompasses multiple dimensions of therapeutic success, such as treatment efficacy, tolerability and safety. The high rates of upadacitinib persistence in our cohort support its efficacy as induction and maintenance therapy in patients with moderate-to-severe UC and CD but also are reassuring in regard to the safety profile of this drug. Those rates are higher than those reported in the pivotal regulatory phase III studies (U-ACHIEVE, U-ACCOMPLISH, U-EXCEL, U-EXCEED, and U-ENDURE)[7,8] and are comparable to those reported in other real-world cohorts. At the end of one year follow-up, the persistence rate for patients with UC was 88.1% in our group of patients. Similarly, recently published data from the PROFUNDUS study reports an 82.8% persistence rate for upadacitinib after 26 weeks of treatment[17]. Regarding CD, Farkas et al[18] showed a 79% probability of treatment persistence with upadacitinib at week 52, a finding consistent with our own observed rate of 72.1%. For CD patients, the persistence rates observed in our cohort exceed those reported in other real-world studies. Recently published data from the Eneida Registry showed a 60% probability of upadacitinib persistence at 12 months in patients with CD[19]. Elford et al[20] reported a 62.8% treatment persistence rate at week 52. Interestingly, in our study, patients with UC achieved significantly higher treatment-persistence rates than those with CD (88.1% vs 72.1%; P = 0.046). Higher upadacitinib persistence rates in UC were also demonstrated in a recently published retrospective cohort study, in which 88% of patients with UC and 78.3% of those with CD remained on upadacitinib at six months of treatment[21]. The higher persistence observed in UC may be explained by the greater phenotypic heterogeneity of CD and may reflect variations in disease biology between colonic and small-intestinal inflammation[22]. Supportive of this hypothesis is our current finding that patients with CD and pure colonic involvement showed better endoscopic outcomes than those with ileal involvement.

The percentages of patients who met our criteria for clinical remission and corticosteroid-free remission after one year of treatment were 63.3% and 51.7%, respectively. Those results align with other real-world studies. In the systematic review and meta-analysis of observational studies by Taxonera et al[23], a clinical remission rate of 64.6% at 24 weeks to 36 weeks was reported for patients with active UC. In the Eneida registry, the clinical remission rate at week 52 exceeded 60% in patients with CD[19]. In another study, Farkas et al[18] showed one-year corticosteroid-free clinical remission rates as high as 86.2% and 66.7% in UC and CD, respectively. The differences observed between studies, may be due to different definitions of clinical remission and sample sizes. Notably, our population predominantly consisted of patients with prior failure to multiple advanced therapies. Our reported rates are much higher than those of the “Ureal study”, as García et al[24] mention a clinical remission rate of 42% at 12 months in IBD patients with a median of four prior advanced therapies. As with persistence, in our cohort, clinical remission rates at one year were numerically higher in patients with UC (74.2%) than in those with CD (51.7%), but not to a statistically significant level. Recently, the STRIDE initiative has emphasized the need to achieve more demanding treatment targets beyond clinical remission alone, as this may be associated with better long-term outcomes and changes in the natural history of IBD[5]. Along that line, it is of great significance that, in the present study, endoscopic remission was achieved in nearly two-thirds of patients with UC (65.2%) and 41.2% of those with CD. Such results may be predictive of a favorable long-term trajectory of disease in patients who receive upadacitinib.

One of the most interesting findings of our study was the association between disease location and endoscopic outcomes in CD. Endoscopic remission was achieved by all patients with isolated colonic disease (4/4), compared to only a minority of those with ileal or ileocolonic disease (3/13). However, given the small number of patients, especially those with colonic disease, these observations should be interpreted with extreme caution and be considered purely hypothesis-generating. Although these findings align with recent research suggesting that ileal and colonic CD may represent partially separate entities, as they demonstrate differences in genetic profiles, immunological pathways, gut dysbiosis, and efficacy of medical treatments[22,25,26], it remains uncertain whether selective JAK1 inhibition is predominantly efficient in colonic CD. Larger, prospective multicenter cohorts are strictly required to investigate this potential association.

An important contribution of our study is the identification of prognostic factors for treatment outcomes. Among patients with UC, smoking and higher baseline CRP levels were independently associated with earlier treatment discontinuation. Smoking has been identified as a negative prognostic factor for clinical remission in IBD patients in the “Ureal study”, in which non-smokers demonstrated a 4.4-fold higher likelihood of achieving clinical remission at 12 weeks compared to smokers[24]. Even though smoking has traditionally been considered as a protective factor for UC[27-29], these findings, alongside the well-established correlations of smoking with MACE, thromboembolic events and malignancy[30-32] support the need for smoking cessation in patients receiving upadacitinib. The association of elevated baseline CRP levels with poorer persistence rates may be indicative of a higher inflammatory burden that makes disease more refractory to treatment. When treatment failure was more broadly defined as treatment discontinuation or the need for treatment intensification, our analysis depicted male sex and extensive disease (Montreal classification E3) as independent predictors in patients with UC. This association is biologically plausible, as extensive disease has systematically been associated with poorer disease prognosis, increased hospitalization rates, requirement of treatment optimization and UC complications[33-35]. To our knowledge, the depicted association between male sex and higher risk for treatment discontinuation or intensification is reported for the first time. While sex-based differences in immune response or hormonal pathways are often suggested as potential mechanisms in autoimmune diseases, such biological explanation remains entirely hypothetical in the context of our small sample size. Finally, we identified female sex and a diagnosis of UC as independent predictors of endoscopic remission. The superior endoscopic outcomes observed in UC are in line with the aforementioned higher treatment persistence and clinical remission rates in our population and are consistent with the better endoscopic outcomes that were seen in UC rather than CD in phase III Randomised Controlled Trials (RCTs)[7,8]. However, whether sex plays a true prognostic role in upadacitinib response requires rigorous evaluation in larger real-world cohorts.

AEs occurred in approximately half of patients in our cohort, with infections being the most frequent. In fact, 21 patients (24.1%) reported at least one infection during the follow-up period. Other frequent AEs included acne, hyperlipidemia and elevated creatine phosphokinase levels, which have all been reported in relation to treatment with upadacitinib. AEs led to treatment discontinuation in 5.7% of our population and hospitalization was required in four patients (4.6%). It is worth noting that the rates of AEs are higher than those reported in other real-world studies[18-20,24,36-38]. This may be due to the very close and prospective observation of patients and the systematic reporting of every possible AE. On the other hand, a systematic review and meta-analysis by Zheng et al[39] reported a pooled AEs rate of 69% in patients with IBD. It is important to note that a definitive causal association between all AEs reported in our study and upadacitinib treatment cannot be established. Regarding infections, which were the most common AEs, every single event was recorded, even if the patient demonstrated mild symptoms. Importantly, no MACE, thromboembolic events, or deaths occurred during follow-up. The majority of AEs occurred during the first 20 weeks of treatment, particularly acne, alterations in lipid profile parameters, and elevations in transaminases. After this initial period, the distribution of most AEs decreased markedly, except for infections, which demonstrated a more dispersed temporal pattern in new-onset events during the follow-up period.

Our study has several strengths. It reflects real-world treatment outcomes of upadacitinib in a predominantly biologic-experienced IBD population. Such cases are often underrepresented in clinical trials, although they constitute a large proportion of patients treated in tertiary centers. Moreover, the median 66-week follow-up period is longer than in most available real-world studies evaluating the efficacy of upadacitinib in IBD, thus allowing for the assessment of more durable treatment outcomes. Notably, the present study identified possible predictors of clinical and endoscopic remission in patients with IBD. Finally, both clinical and objective endoscopic outcomes were evaluated.

Certain important limitations should also be acknowledged. Our study incorporated a single-ethnic cohort, which may restrict the generalizability of our results to ethnically and geographically diverse IBD populations with varying genetic backgrounds. Second, the observational design lacked a control group, preventing firm causal inferences or comparative effectiveness evaluations. Endoscopic assessment after upadacitinib initiation was available only in a fraction of patients, introducing a potential selection bias. Our sample size was relatively small, particularly for subgroup and multivariable analyses, and no biochemical outcomes were reported. Finally, our study population consisted of patients failing one or more lines of advanced therapies. Consequently, the scope of applicability of our study conclusions is primarily restricted to real-world, bio-experienced IBD patients treated in tertiary care settings. These findings provide preliminary real-world evidence that requires confirmation in larger, prospective, multi-center, controlled cohorts.

CONCLUSION

In our real-world cohort of patients with IBD refractory to multiple previous advanced therapies, upadacitinib demonstrated high treatment persistence rates at one year, promising clinical and endoscopic effectiveness, alongside an acceptable safety profile. Treatment outcomes appeared particularly favorable for UC patients, where both persistence and endoscopic remission rates were higher than those observed in CD. The findings of the present study support the use of upadacitinib as an effective and safe therapeutic option for patients with difficult-to-treat IBD. Findings regarding specific prognostic factors should be interpreted with caution, given the limited sample sizes for subgroup and multivariable analyses, and require further validation in larger, prospective multicenter cohorts.