Long-term Extension Research Articles

Long-term effectiveness of ARRY-371797 in people with dilated cardiomyopathy and a faulty LMNA gene: a plain language summary.

This summary explains the results of a long-term extension study on the effects of a specific medicine. A long-term extension study allows people who have already completed a research study to continue taking treatment. Researchers can then look at how a treatment works over a long period of time. This extension study looked at the effects of a medicine called ARRY-371797 (also known as PF-07265803) in people with dilated cardiomyopathy (DCM for short) caused by a faulty lamin A/C gene (also known as the LMNA gene). This condition is called LMNA-related DCM. In people with LMNA-related DCM, the heart muscle becomes thinner and weaker than normal. This can lead to heart failure, where the heart is unable to pump enough blood around the body. The extension study allowed people who had completed an earlier 48-week study to continue taking ARRY-371797 for another 96weeks (around 22months). 8 people joined the extension study and continued with the dose of ARRY-371797 that they had taken in the first study. This means that people could have taken ARRY-371797 continuously for up to 144weeks (around 2years and 9months). Using the 6-minute walk test (6MWT for short), researchers regularly checked people taking ARRY-371797 to see how far they could walk. Throughout the extension study, people were able to walk further than they could before they started taking ARRY-371797. This suggests that people could maintain the improvements in their ability to do daily activities with long-term ARRY-371797 treatment. Researchers also looked at how severe people's heart failure was by using a test that measures levels of a biomarker called NT-proBNP. A biomarker is something found in the body that can be measured to indicate the extent of a disease. Throughout this study, the levels of NT-proBNP in people's blood was lower than before they started taking ARRY-371797. This suggests that they maintained stable heart function. Using the Kansas City Cardiomyopathy Questionnaire (KCCQ for short), researchers asked people about their quality of life, and if they experienced any side effects. A side effect is something that people feel while taking a treatment. Researchers evaluate if a side effect is related to the treatment or not. Some improvement in KCCQ response during the study was seen, although results were varied. There were no serious side effects that were considered related to treatment with ARRY-371797. Researchers found that the improvements in functional capacity and heart function seen with ARRY-371797 treatment in the original study were maintained with long-term treatment. Larger studies are needed to determine if ARRY-371797 could be an effective treatment for people with LMNA-related DCM. One such study (called REALM-DCM) was started in 2018 but ended early, as it was unlikely to show a clear treatment benefit of ARRY-371797. Phase 2 long-term extension study (NCT02351856) Phase 2 study (NCT02057341) Phase 3 REALM-DCM study (NCT03439514).

The Late Devonian contraction along the Precaspian Basin north margin: Transpressional range building in the intracratonic domain

The evolution of intraplate sedimentary basins located in the vicinity of an active convergent plate boundary is often controlled by the collisional dynamics of the adjacent orogen. The transfer of compression from the orogen to the platform's interior results in the formation of complex structural geometry and kinematics that often reactivate older crustal faults, focus far-field stresses and control the evolution of associated sedimentary basins. One place where this localisation can be optimally understood is the Precaspian Basin, situated at the SE periphery of the East European Craton and bordered to the east by the Uralian orogen. The Precaspian Basin and its northern margin experienced long-term extension and subsidence interrupted by several short-lived shortening episodes. To understand the impact and role of pre-existing basement structures on the geometry and kinematics of the subsequent deformation, we analysed subsurface data from the northern margin of the Precaspian Basin by the means of 3D seismic interpretation correlated with wells and structural modelling. The analysis results provide new insights into the kinematic effects of the Late Devonian contraction event. The superposition of stratigraphic units above and below the angular unconformity suggests an intra-Famennian age of the deformation event. Steeply dipping bi-verging reverse fault zones associated with variable amounts of sinistral strike-slip movement component display an arcuate geometry, trending from WSW-ENE to NW-SE. The distribution of deformation indicates that this complex kinematic pattern was driven by a NE-SW oriented contraction and transpression, where faults show the characteristics of a restraining bend area. This area is interpreted as part of a regional transcurrent fault system developed on the northern periphery of the Precaspian Basin. Furthermore, the study results suggest that intralithospheric stress localisation transmitted by the Paleouralian subduction zone resulted in the reactivation of pre-existing basement structures, propagation of faults, and localized short-term exhumation of the Precaspian Basin north margin. Generated by the collision of Magnitogorsk volcanic arc with East European Craton, far-field stress transfer produced a zone of oblique deformation. These data and interpretation demonstrate that the northern margin of the Precaspian Basin is an excellent natural setting to investigate and better understand mechanisms of far-field strain localisation and reactivation of deformational structures in stable platform areas resulting in the intracratonic mountain building process.

Population pharmacokinetics of duloxetine in Japanese pediatric patients with major depressive disorder

The objectives of this analysis were to characterize the pharmacokinetics of duloxetine in Japanese pediatric patients aged 9–17 years with major depressive disorder (MDD) and to explore potential intrinsic factors affecting its pharmacokinetics. A population pharmacokinetic (PK) model was developed with plasma steady-state duloxetine concentrations from Japanese pediatric patients with MDD in an open-label long-term extension trial in Japan (ClinicalTrials.gov Identifier: NCT03395353). Duloxetine pharmacokinetics in Japanese pediatric patients was well described by a one-compartment model with first-order absorption. The population mean estimates of CL/F and V/F of duloxetine were 81.4 L/h and 1170 L, respectively. Patient intrinsic factors were assessed for their potential influence on duloxetine apparent clearance (CL/F). Only sex was identified as a statistically significant covariate of duloxetine CL/F. Duloxetine pharmacokinetic parameters and model-predicted duloxetine concentrations at steady state in the Japanese pediatric population were compared with those in Japanese adults. The mean duloxetine CL/F in pediatrics is slightly higher than adults, it is, however, expected that comparable steady-state duloxetine exposure in pediatric patients can be achieved with the approved dose regimen for adults. The population PK model provides useful information to understand the pharmacokinetic characteristics of duloxetine for Japanese pediatric patients with MDD. ClinicalTrials.gov identifierNCT03395353

Pooled Analysis of Baricitinib Tolerability in Patients With Atopic Dermatitis in Relation to Acne, Headache, and Gastrointestinal Events From 8 Clinical Trials.

Background: Tolerability issues including acne, nausea, and headache have been reported with Janus kinase (JAK) inhibitors for moderate-to-severe atopic dermatitis (AD). Objectives: To report outcomes of tolerability adverse events (AEs) for baricitinib, a JAK1/JAK2 inhibitor, in patients with moderate-to-severe AD. Methods: Acne, headache, and gastrointestinal AEs are reported from placebo-controlled and long-term extensions of pooled data in the baricitinib AD clinical trial program. Proportions of patients with AEs, incidence rates (IRs)/100 patient-years at risk, and median time to onset/duration of AEs were calculated. Results: In 2531 patients treated with baricitinib, most AEs were mild to moderate in severity. Headache was the most common AE of tolerability (median of 14-26 days after first dose of baricitinib, lasting ≤3 days). IRs of acne were <5 in any group lasting up to a median of 90 days with no severe AEs. Diarrhea was the most common gastrointestinal AE, lasting a median of ≤7 days. There were few study drug interruptions (n = 6) and permanent discontinuations (n = 5) for tolerability AEs. Conclusions: For the AEs of tolerability analyzed, baricitinib appears to be well tolerated. Overall, the frequency of these AEs in patients being treated for moderate-to-severe AD was low with few leading to study drug interruption or permanent discontinuation. Clinical Trial Registration number: NCT02576938; NCT03334396; NCT03334422; NCT03428100; NCT03435081; NCT03733301; NCT03334435; NCT03559270.

Suprachoroidal triamcinolone acetonide injectable suspension for macular edema associated with noninfectious uveitis: an in-depth look at efficacy and safety.

Patients with macular edema (ME) associated with uveitis (UME) are at risk for vision loss and decreased quality of life, and they often experience high health care costs and rates of workforce absenteeism. Systemically or locally delivered corticosteroids are the mainstay of treatment for UME. Although traditional corticosteroid treatments may demonstrate high levels of efficacy, systemic delivery carries the risk of potentially serious systemic adverse effects (AEs), and standard local modes of delivery may be associated with low bioavailability in posterior ocular tissues and steroid-associated AEs due to anterior ocular tissue exposure. Drug injection into the suprachoroidal space (SCS) allows for targeted delivery to chorioretinal tissues with high bioavailability in the posterior segment, as well as for inherent drug sequestration away from the anterior segment, which may lower the risk of AEs associated with anterior tissue exposure to steroids. A novel triamcinolone acetonide (TA) injectable suspension formulated for administration to the SCS, SCS-TA (Xipere®; Bausch + Lomb), received FDA approval in 2021 for the treatment of UME. It is administered via the SCS Microinjector® (Clearside Biomedical, Inc), a device specifically designed for SCS delivery of ocular therapeutics. This approval was based on results from the phase 3 PEACHTREE clinical trial (NCT02595398) that demonstrated the clinical efficacy-including significantly increased visual acuity and decreased central subfield thickness-and safety of SCS-TA in patients with UME. Results from this trial, as well as from its long-term observational extension (MAGNOLIA; NCT02952001) and an open-label safety study (AZALEA; NCT03097315), support the possibility that treatment with SCS-TA may address the burden of disease in patients with UME.

Effect of beta-blocker therapy on the response to mavacamten in patients with symptomatic obstructive hypertrophic cardiomyopathy.

In the EXPLORER-HCM trial, mavacamten improved exercise capacity and symptoms in patients with obstructive hypertrophic cardiomyopathy (oHCM). Mavacamten effects on the primary endpoint, a composite of peak oxygen consumption (VO2 ) and New York Heart Association (NYHA) class, were greater in patients not receiving background beta-blockers than in those receiving beta-blockers. We sought to determine if the effect of background treatment was consistent across other clinically meaningful parameters. Subgroup analyses by beta-blocker use were performed in patients with oHCM from the EXPLORER-HCM and mavacamten long-term extension (MAVA-LTE) studies. In EXPLORER-HCM, 189 patients (75.3%) were receiving beta-blockers, and 62 (24.7%) were receiving non-dihydropyridine calcium channel blockers or no background HCM medication; 170 patients (90.4%) receiving beta-blockers had chronotropic incompetence. Improvements in peak VO2 at week 30 with mavacamten versus placebo were lower with beta-blockers (mean difference [95% confidence interval (CI)]: 1.04 [0.12, 1.95] ml/kg/min) than without beta-blockers (mean difference [95% CI]: 2.69 [1.29, 4.09] ml/kg/min); improvements in non-heart rate-dependent parameters (VE /VCO2 slope) appeared unaffected by beta-blockers. Improvements in functional capacity parameters at week 30 with mavacamten versus placebo were independent of beta-blockade for post-exercise left ventricular outflow tract gradient (mean difference [95% CI]: -37.9 [-48.0, -27.9] mmHg with beta-blockers; -33.5 [-53.6, -13.3] mmHg without beta-blockers), proportion of patients with reduction of ≥1 NYHA class, Kansas City Cardiomyopathy Questionnaire clinical summary scores and N-terminal pro-B-type natriuretic peptide. Mavacamten benefits were reproduced and maintained in MAVA-LTE regardless of beta-blockade. Mavacamten improved measures of functional capacity, left ventricular outflow tract obstruction, symptom burden and biomarkers in patients with HCM regardless of beta-blocker use. Beta-blocker use was often associated with chronotropic incompetence, affecting peak VO2 and other heart rate-dependent measures, but had minimal impact on heart rate-independent measures.

Plain Language Summary of Publication of the safety and efficacy of ARRY-371797 in people with dilated cardiomyopathy and a faulty LMNA gene.

This plain language summary describes the results of a study looking at the effects of a medicine called ARRY-371797 (also known as PF-07265803) in people with dilated cardiomyopathy (DCM for short) caused by a faulty LMNA gene. This condition is called LMNA-related DCM. DCM happens when the heart becomes bigger and weaker than normal, impacting functional capacity and leading to symptoms of heart failure. This means the heart is not able to pump blood around the body as easily, and people are unable to do as much in their daily lives (like getting dressed and going shopping). People may inherit a faulty LMNA gene from one of their parents, or a faulty LMNA gene may develop when mistakes happen during cell growth and replication. ARRY-371797 targets a specific mechanism in the body that can lead to heart problems in people with a faulty LMNA gene. As ARRY-371797 is not currently approved for use outside of clinical trials, it doesn't currently have an easily recognizable trade name. 12 American people (average age 50years) with LMNA-related DCM took part in the study and received 400mg or 100mg of ARRY-371797 twice daily for 48weeks. People knew which dose of ARRY-371797 they were taking. People were checked after 4, 12, 24, 36 and 48weeks of taking ARRY-371797 to see how far they could walk in the 6-minute walk test (6MWT for short). The level of NT-proBNP in their blood was also measured. NT-proBNP is a biomarker used to measure the severity of heart failure. A biomarker is something found in the body that can be measured to indicate the extent of a disease. -After taking ARRY-371797 for 12weeks, people were able to walk further in the 6MWT and had lower levels of NT-proBNP in their blood. This suggests improvement in functional capacity (exercise tolerance) and heart function. Researchers also asked people about their quality of life using the Kansas City Cardiomyopathy Questionnaire (KCCQ for short), and looked for any side effects. -Researchers saw some improvement in KCCQ scores. -Researchers saw no major side effects that they considered to be related to ARRY-371797 treatment. A side effect is something that people feel was caused by a medicine or treatment. Overall, this study showed that people with LMNA-related DCM who took ARRY-371797 had improved functional capacity (exercise tolerance), improved heart function, and improved quality of life. Phase 2 study (NCT02057341) Phase 2 long-term extension study (NCT02351856) Phase 3 REALM-DCM study (NCT03439514).

P438 Efficacy and safety of ustekinumab for Ulcerative Colitis through 4 years in Asian patients: Final results from the sub-population of UNIFI long-term extension

Abstract Background Ustekinumab (UST) is an interleukin-12/23 p40 antagonist approved for the treatment of moderately to severely active ulcerative colitis (UC). Here, we report the final efficacy and safety results of UST in the Asian subpopulation in the UNIFI long-term extension (LTE) through 4 years (Week [WK] 200). Methods Overall, 87 intravenous (IV) UST induction WK8 responders were randomised to subcutaneous (SC) maintenance therapy: 34 placebo (PBO); 24 UST 90 mg every 12 weeks (q12w); 29 UST 90 mg q8w. The nonrandomised patients (pts) included 22 IV UST induction nonresponders who received SC UST at WK8, responded at WK16, and continued on SC UST q8w; and 12 responders to IV PBO induction who received SC PBO.1,2 Pts who completed WK44 continued on SC UST q8w or PBO treatment in the LTE, and PBO pts were discontinued after unblinding. Starting at WK56, randomised pts with UC worsening could adjust to SC UST 90 mg q8w. Efficacy was evaluated in UST-randomised pts (n=53) using symptomatic remission (Mayo stool frequency subscore of 0 or 1 and rectal bleeding subscore of 0). Safety was evaluated for all pts treated in the LTE, including randomised and nonrandomised populations until WK220. Results Among all UST-randomised pts at maintenance baseline WK0 (intent-to-treat population with nonresponder imputation for missing data and treatment failure criteria), 49.1% were in symptomatic remission at WK200 (biologic naïve 77.8.%; failures 34.4%); 49.1% of pts were in corticosteroid-free symptomatic remission at WK200 (Table 1; Fig. 1-2). Overall, 43.5% of biologic failure and 0% of biologic naïve pts who were randomised to UST and treated in the LTE discontinued treatment between WK44 and WK200. Among randomised pts who continued UST in the LTE, 63.4% were in symptomatic remission at WK200; 78.6% of pts with observed data at WK200 were in symptomatic remission (Table 1; Fig. 3); and 80.0% of those in clinical remission at WK44 were in symptomatic remission at WK200. Safety events were similar among UST-treated pts compared with PBO. From maintenance baseline WK0 through WK220, pts receiving UST (combined UST) and PBO had 241.8 and 56.5 total pt-years of follow-up, respectively. Safety events per 100 pt-years of follow-up for combined UST vs PBO were AEs: 222.48 vs 320.29, SAEs: 11.58 vs 7.08, and serious infections: 1.65 vs 1.77. During the final year of the LTE, no deaths or major cardiovascular events were reported in UST-treated Asian pts. Conclusion Consistent with the overall study population in UNIFI LTE1,2, Asian pts receiving SC UST generally maintained clinical benefit through 4 years. No new safety signals were observed. 1Afif W, et al. UEGW. Vienna, Austria, October 8-11, 2022. 2Sands BE, et al. ACG. Charlotte, NC, October 21-26, 2022

OP35 Efficacy and safety outcomes up to ~4 years of treatment with filgotinib 200 mg among patients with Ulcerative Colitis: Results from the SELECTIONLTE study

Abstract Background Filgotinib (FIL) is a once-daily, oral, Janus kinase 1 preferential inhibitor approved for the treatment of ulcerative colitis (UC). FIL 200 mg (FIL200) was effective in inducing and maintaining clinical remission vs placebo (PBO) and well tolerated in patients with UC in the randomized, phase 2b/3 SELECTION trial (NCT02914522).1 The efficacy and safety of continued FIL200 therapy are being assessed in the continuing SELECTION long-term extension (LTE) study. Methods The SELECTION programme comprises SELECTION and the associated LTE study (SELECTIONLTE; NCT02914535). In SELECTION, adults with moderately to severely active UC received induction (IND) FIL200, FIL 100 mg (FIL100) or PBO once daily for 11 weeks. Patients in clinical remission or with a Mayo Clinic Score (MCS) response at week 10 (responders) could enter the 47-week Maintenance (MNT) Study. SELECTIONLTE enrolled patients who completed IND and MNT (completers), patients who were not responders at IND week 10 (non-responders), and patients with disease worsening during MNT. This interim analysis assessed the efficacy and safety of open-label FIL200 through LTE week 144 in completers and LTE week 192 in non-responders, respectively, representing a total of 3.9 years of treatment each (completers: 58 + 144 weeks; non-responders 10 + 192 weeks). The cut-off date was 24 February 2022. Continuous measures of partial MCS (pMCS) and proportions of patients achieving the minimal clinically important difference (MCID) in total IBDQ score (≥16-point increase vs baseline) were assessed without imputation (as observed). Safety was assessed by the exposure-adjusted incidence rate per 100 patient-years of exposure (PYE) for treatment-emergent adverse events (TEAEs) and TEAEs of interest. Results This analysis included 148 completers and 372 non-responders (IND FIL100, n=212; IND FIL200, n=160). Among completers, the reductions in mean pMCS in SELECTION over weeks 0–10 were maintained up to LTE week 144 (Figure 1). In non-responders, mean pMCS decreased from LTE baseline to week 192 (Figure 2). High proportions of completers (&amp;gt;90%) and non-responders (&amp;gt;60%) achieved the MCID in IBDQ score; proportions were maintained over time (Table 1). Safety events among all patients treated with open-label FIL200 in the LTE study (total PYE=2055.5) are reported in Table 2. Conclusion: FIL200 was effective in maintaining symptomatic remission and health-related quality of life for up to ~4 years. No new safety signals were observed over ~4 years of treatment, and safety events were comparable with those seen in previous analyses.2 1. Feagan BG et al. Lancet 2021;397:2372–84. 2. Feagan BG et al. J Crohns Colitis 2022;16(Suppl1):i456–7.

OP15 Efficacy of ustekinumab for Ulcerative Colitis through 4 years: Final clinical and endoscopy outcomes from the UNIFI long-term extension

Abstract Background Ustekinumab (UST) is an interleukin12/23p40 antagonist approved for treatment of moderate to severe Ulcerative Colitis (UC). Here we report final clinical outcomes based on the Mayo score, including the endoscopy subscore, from the UNIFI long-term extension (LTE) study through 4 years of UST treatment. Methods Overall, 523 intravenous UST induction responders were randomised to subcutaneous maintenance therapy: 175 placebo (PBO); 172 UST 90 mg every 12 weeks (q12w); 176 UST 90 mg q8w. A total of 284 UST patients (pts) completed week (wk)44 and continued treatment in the LTE; pts receiving PBO were discontinued after study unblinding. Starting at wk56, randomised pts with UC worsening could receive a dose adjustment: PBO to UST q8w, UST q12w to UST q8w, and UST q8w to UST q8w (sham adjustment). Outcomes based on the Mayo score (including endoscopy assessed by a local reader) were evaluated at the final efficacy visit (wk200); pts who had treatment failure (i.e., had ostomy or colectomy or discontinued UST due to lack of therapeutic effect or worsening UC) before wk200 were also included, and they were imputed as nonresponders. Clinical remission was defined as a Mayo score ≤2 points and no individual subscore &amp;gt;1; clinical response as a decrease in Mayo score of ≥30% and ≥3 points from induction baseline with either a decrease in rectal bleeding subscore of ≥1 from induction baseline or a rectal bleeding subscore of 0 or 1; modified Mayo score (without Physician’s Global Assessment subscore) response as a decrease in modified Mayo score of ≥30% and ≥2 points from induction baseline with either a decrease in rectal bleeding subscore of ≥1 from induction baseline or a rectal bleeding subscore of 0 or 1; and endoscopic improvement as an endoscopy subscore of 0 or 1. Results Among 205 pts who were randomised to UST at maintenance baseline and continued treatment in the LTE, who either had Mayo score data (including endoscopy) at wk200 or had experienced treatment failure, 58.0% were in clinical remission, 80.0% were in clinical response, 79.5% were in modified Mayo score response, and 67.3% showed endoscopic improvement (Figure 1). Using an alternative “as observed case” analysis (without imputation for treatment failures), among 171 randomised pts who continued UST treatment through LTE and had an endoscopy score at wk200, 69.6% were in clinical remission, 95.9% were in clinical response, 95.3% were in modified Mayo score response, and 80.7% showed endoscopic improvement. Conclusion Among these pts with a history of moderate to severe UC who continued UST treatment, approximately 2/3 or more were in clinical remission, clinical response, and/or demonstrated endoscopic improvement at 4 years.

P671 Ustekinumab in paediatric patients with moderately to severely active Crohn's disease: Results from the UniStar study long-term extension

Abstract Background Ustekinumab (UST) is an approved biologic for the treatment of adults with Crohn’s disease (CD). Through 16 weeks (wks), UST demonstrated high rates of efficacy, low rates of immunogenicity, and safety in paediatric patients (pts) similar to that observed in adults.1 Here, we evaluated UST in paediatric CD pts through 268 wks. Methods UniStar was a phase 1, multicentre, double-blind study in paediatric pts (2-18 years old; body weight ≥10kg) with moderately to severely active CD. Pts (n=45) were randomised 1:1 and stratified by body weight (&amp;lt;40kg or ≥40kg) to low (3 mg/kg [&amp;lt;40kg]; 130 mg [≥40kg]) or high (9 mg/kg [&amp;lt;40kg]; 390 mg [≥40kg]) dose intravenous UST. At wk 8, pts received 1 subcutaneous maintenance dose (2 mg/kg [&amp;lt;40kg]; 90 mg [≥40kg]). Pts responding to UST at wk 16 were eligible to enter the long-term extension (LTE) and continued UST maintenance dosing every 8 wks up to wk 268. This study reports efficacy, safety, pharmacokinetics (PK), and immunogenicity evaluated at wks 16, 24, 48, 224, and through 240 wks, as applicable. The primary visit was wk 48, representing outcomes after ~1 year of UST treatment. Results Of the 44 pts treated with UST, 34 (77%) responded and entered the LTE; the number of pts receiving treatment at wks 48, 104, 152, and 208 was 26 (77%), 16 (47%), 12 (35%), and 8 (24%), respectively. Among pts in the LTE (n=34), baseline median age (range) was 13.0 (6.0-17.0) years, 62% were female, 47% weighed &amp;lt;40kg, 18% weighed &amp;lt;30kg, and 94% previously failed anti-tumor necrosis factor therapy. At wk 48, 41% (14/34) of pts achieved clinical remission, and 59% (20/34) achieved clinical response. The proportion of pts with abnormal C-reactive protein (CRP) levels at baseline was 71% (24/34), and 59% (16/27) of pts achieved normalization of CRP (&amp;lt;3mg/L) at wk 48. By the final safety visit at wk 240, 91% of pts reported ≥1 adverse event (AE), and 15% discontinued due to AEs, with CD worsening as the most frequent reason. Infections occurred in 74% of pts, and 32% had serious AEs. The majority of serious AEs were gastrointestinal disorders associated with CD. No injection-site reactions, serious infections, malignancies, or deaths were reported. Median serum UST concentrations tended to be lower in pts weighing &amp;lt;40kg compared with pts weighing ≥40kg, but were generally consistent from wks 16 to 268 and remained detectable through wk 200 (Figure). Incidence of UST antibodies was low (1/34; 3%). Conclusion In the UniStar LTE, the efficacy, safety, PK, and immunogenicity of UST in paediatric pts with CD was generally consistent with adult pts for as long as 4 years of treatment. Reference: 1. Rosh JR et al. J Crohns Colitis. 2021;15(11):1931.

P520 Thromboembolic and major adverse cardiovascular events among patients in the filgotinib clinical trial programme

Abstract Background The once-daily, oral, Janus kinase 1 preferential inhibitor filgotinib (FIL) is approved for the treatment of moderately to severely active rheumatoid arthritis (RA) and ulcerative colitis (UC) in the UK, the EU and Japan.1 To further understand the safety profile of FIL across indications, we evaluated the risk of major adverse cardiovascular events (MACEs) and venous thromboembolic events (VTEs) in patients treated with FIL 200 mg (FIL200) or FIL 100 mg (FIL100). Methods An integrated analysis was conducted with RA data from five phase 2/3 trials and two long-term extension (LTE) trials of FIL, and UC data from two phase 2/3 trials and one LTE trial of FIL (herein termed ‘overall’ data). Subgroup analyses by age and cardiovascular (CV) risk factors (excluding age) were conducted. Real-world data were extracted from a systematic literature review. Exposure-adjusted incidence rates (EAIRs) or incidence rates per 100 patient-years of exposure and 95% confidence intervals (CIs) were estimated for MACEs and VTEs. For clinical trial data, MACEs and VTEs included positively adjudicated events only. Results In pooled data, baseline characteristics were generally similar across treatment arms for RA and UC (Tables 1 and 2). In the RA data, MACE EAIRs were 0.29 and 0.41 for the overall FIL200 and FIL100 arms, respectively (Figure 1). MACE IRs were 0.45–0.77 for the general RA population. Higher MACE EAIRs were reported in the subgroup of patients with RA aged ≥65 years vs &amp;lt;65 years. In the RA subgroup with CV risk, MACE EAIRs were 0.53 for FIL200 and 0.64 for FIL100. VTE rates were 0.19 for both FIL200 and FIL100 overall in patients with RA, and 0.32–0.44 for the general RA population. Numerically higher VTE EAIRs with overlapping CIs were seen in patients with RA aged ≥65 years vs &amp;lt;65 years, and in those with vs without CV risk. In the UC data, MACE rates were 0.29, 0.35 and 0.87 in the overall FIL200 and FIL100 arms and in the general UC population, respectively (Figure 2). VTE rates were 0.08, 0.18 and 0.39 in the overall FIL200 and FIL100 arms and in the general UC population, respectively. For both MACE and VTE EAIRs, wide CIs were reported in patients with UC aged ≥65 years and in those with CV risk, owing to low numbers of events and/or patients in these subgroups. Conclusion No association was identified between FIL200 treatment and an increased risk of MACEs or VTEs compared with the general RA or UC population. Patients aged ≥65 years or with CV risk had slightly higher rates of MACEs and VTEs than those in other subgroups; however, overlapping CIs suggested no real difference. Further work should examine real-world data from FIL-treated patients and longer follow-up. 1. https://www.ema.europa.eu/en/medicines/human/EPAR/jyseleca

DOP45 Predicting late responders to filgotinib in patients with moderately to severely active Ulcerative Colitis: A post hoc analysis of SELECTION and SELECTIONLTE

Abstract Background Filgotinib (FIL) is a once-daily, oral, Janus kinase 1 preferential inhibitor approved for the treatment of ulcerative colitis (UC) in Europe and Japan. Previous analyses of the SELECTION programme showed that, in patients with moderately to severely active UC, C-reactive protein, rectal bleeding and stool frequency at day 7 predicted response to FIL at week 10,1 and that patients who did not respond to FIL induction benefited from an additional 12 weeks of FIL 200 mg.2 This post hoc analysis was conducted to identify factors associated with late response to FIL. Methods The SELECTION programme comprises a phase 2b/3, randomized, double-blind, placebo (PBO)-controlled study (SELECTION; NCT02914522) and its open-label, long-term extension (SELECTIONLTE; NCT02914535). Adult patients aged 18–75 years were randomized to FIL 100 mg, FIL 200 mg or PBO once daily for 11 weeks in the induction study. This post hoc analysis only included patients receiving FIL 200 mg who did not have clinical remission or a Mayo Clinic Score (MCS) response at week 10 of induction and who subsequently entered SELECTIONLTE on FIL 200 mg. Patient characteristics at induction baseline and changes in surrogate markers of response from baseline to week 4 of induction were compared between patients who had observed partial MCS (pMCS) response (pMCS late responders) or observed non-response (pMCS non-responders) at week 12 of SELECTIONLTE using univariate and multivariate logistic regression models. Results In total, 160 patients were included (102 [63.8%] pMCS late responders and 58 [36.3%] pMCS non-responders) (Table 1). In univariate analysis, prior use of biologics was not significantly associated with late response to FIL. Changes in surrogate markers of response showed that none were significantly associated with late response (Figure 1). In multivariate analysis of baseline characteristics, older age was positively associated with late response to FIL (odds ratio [OR]: 1.042, 95% confidence interval [CI]: 1.011–1.073, p = 0.007), while body mass index (BMI) &amp;lt; 18.5 kg/m2 (OR: 0.041, 95% CI: 0.003–0.519, p = 0.014) was negatively associated with late response. Conclusion In SELECTION programme participants, older age was positively associated with late response to FIL 200 mg and BMI &amp;lt; 18.5 kg/m2 was negatively associated with late response. These factors should be taken into account when monitoring patients on FIL with insufficient response at week 10. No other strong predictors were identified. The results of this analysis should be interpreted in the context of previous findings.1 1. Feagan B et al. United European Gastroenterol J 2022;10(S8):462–3 2. Vermeire S et al. United European Gastroenterol J 2021;9(S8):215–6

P518 Safety of upadacitinib in ulcerative colitis: Long-term data from the phase 3 open-label extension study (U-ACTIVATE)

Abstract Background Upadacitinib (UPA), an oral reversible and selective JAK inhibitor, has demonstrated efficacy, with a well-characterised safety profile, in two replicate induction studies of UPA 45 mg once daily (QD; U-ACHIEVE induction and U-ACCOMPLISH), and a maintenance study of UPA 15 mg or UPA 30 mg QD (U-ACHIEVE maintenance).1 Here, we report safety data from the ongoing, Phase 3, open-label, 288-week long-term extension (LTE) study in patients with ulcerative colitis (UC) who previously participated in the induction studies. Methods Rates of treatment-emergent adverse events (TEAEs) and TEAEs of special interest were calculated by treatment sequence in patients who received ≥1 dose of UPA in the induction studies, maintenance study, or by the data cut-off of 13 December 2021 in the LTE study. Rates are presented as exposure-adjusted event rates (EAERs; events/100 patient-years [PYs]), except for events that typically occur as a single instance, which are presented as exposure-adjusted incidence rates (EAIRs; n/100 PYs). PYs were calculated from the first dose of UPA 45 mg received during induction through to the last UPA dose received or the data cut-off date for ongoing patients. Results Overall, 1,308 patients representing 2,350.2 PYs of UPA exposure were evaluated (mean age: 43 years; female: 38%). Across treatment groups, EAERs of serious TEAEs, severe TEAEs, TEAEs leading to UPA discontinuation, TEAEs with reasonable possibility of being related to UPA, or TEAEs leading to death were numerically similar across all UPA treatment sequences (Table 1). EAERs of serious infections, herpes zoster, neutropenia, hepatic disorder, and creatine phosphokinase elevation were numerically higher with UPA 15 mg and 30 mg vs placebo; and creatine phosphokinase elevation with UPA 30 mg vs UPA 15 mg (Table 2). The EAER of anaemia was numerically lower with UPA 30 mg vs UPA 15 mg and placebo. Of note, no remarkable differences were observed in EAIRs of adjudicated major adverse cardiovascular events, adjudicated venous thromboembolic events, or malignancy (excluding non-melanoma skin cancer). Conclusion: Long-term use of UPA 15 or 30 mg among patients with UC continuing to benefit from prolonged therapy representing 2,350.2 PYs of exposure was well tolerated. The ongoing safety profile of UPA is consistent with previous analyses, with no new safety signals identified compared with previous UC analyses following 52 weeks’ maintenance therapy,1,2 or other immune-mediated indications for which UPA is approved.3,4

DOP39 Endoscopic, histologic, and combined outcomes with reinitiation of ozanimod after temporary discontinuation: 2-year interim analysis of the True North open-label extension study

Abstract Background Mucosal healing, evaluated by a combination of endoscopic and histologic examinations, is a desired outcome in ulcerative colitis (UC), but few long-term or open-label extension (OLE) studies of UC treatments have assessed these outcomes. In the phase 3 True North (TN) study, ozanimod 0.92 mg once daily demonstrated efficacy and safety for up to 52 weeks in patients with moderately to severely active UC. The current analysis evaluated long-term endoscopic and histologic endpoints in patients who reinitiated ozanimod after experiencing disease relapse when rerandomised to placebo during the TN maintenance period (MP). Methods Responders to ozanimod 10-week induction therapy were rerandomised 1:1 to ozanimod (n=230) or placebo (n=227); this OLE interim analysis (data cutoff: 10 January 2022, the point at which the OLE Week 94 disposition is available for patients) included all patients who relapsed on placebo during the MP and reinitiated ozanimod in the OLE. Endpoints included percentages of patients with endoscopic improvement (endoscopic subscore of ≤1 point), histologic remission (Geboes index score &amp;lt;2), and mucosal healing (endoscopy subscore of ≤1 point and Geboes index score &amp;lt;2) at OLE Weeks 46 and 94. Data were evaluated as observed case (OC) and nonresponder imputation (NRI) analyses. Results Seventy-seven (34%) patients rerandomised to placebo experienced a disease relapse in the TN MP. Median time to relapse was 12.7 weeks. At data cutoff, all 77 patients had either completed OLE Week 94 (51.9%) or discontinued treatment. At OLE Week 46, endoscopic improvement, histologic remission, and mucosal healing were achieved by 50%, 56%, and 42% of patients, respectively (OC analysis). Proportions with these endpoints further increased to 60%, 68%, and 48%, respectively, by OLE Week 94. Per NRI analysis, endoscopic improvement, histologic remission, and mucosal healing at OLE Weeks 46 and 94 were observed in 30% and 27%, 31% and 25%, and 23% and 17% of patients, respectively (Figure). Furthermore, 52% of patients with endoscopic improvement at OLE Week 46, 58% of patients with histologic remission at OLE Week 46, and 44% of patients with mucosal healing at OLE Week 46 continued to sustain the respective endpoints at OLE Week 94 with another year of ozanimod treatment. Conclusion In patients who experienced disease relapse after rerandomisation to placebo, reinitiation of ozanimod was associated with recapture of response and long-term endoscopic and histologic benefits, including histologic remission and mucosal healing. These data suggest that patients who experience disease relapse when ozanimod is temporarily discontinued may benefit from resumption of ozanimod treatment.

P781 Treatment outcome of tofacitinib dose reduction to 5 mg BID vs remaining on 10 mg BID in patients with Ulcerative Colitis who were in stable remission on 10 mg BID: Updated 30-month data from the double-blind, randomised RIVETING study

Abstract Background Tofacitinib is an oral small molecule Janus kinase inhibitor for the treatment of ulcerative colitis. The long-term, Phase 3b/4 RIVETING study (NCT03281304) assessed the efficacy and safety of tofacitinib dose reduction from tofacitinib 10 mg twice daily (BID) maintenance therapy to 5 mg BID in patients (pts) in stable remission.1 We present a final analysis of the RIVETING study after ≥30 months of treatment. Methods RIVETING was a double-blind, randomised, parallel-group study with a 42-month (M) duration. Eligible pts had received tofacitinib 10 mg BID for ≥2 consecutive years in an open-label, long-term extension study (NCT01470612), had been in stable remission for ≥6 months and corticosteroid-free for ≥4 weeks prior to baseline. The primary efficacy endpoint was remission at M6 based on modified Mayo (mMayo) score (endoscopic and stool frequency subscores of ≤1 and rectal bleeding subscore of 0).1 RIVETING was terminated (primary objective met) when all pts had passed M30 (or discontinued); some pts had already completed all visits up to M42. Here, we assess efficacy at M30 and safety throughout. Results Overall, 140 pts were randomised (1:1) to tofacitinib 5 or 10 mg BID; 50.0% and 62.9% were in remission based on mMayo score at M30 in the 5 and 10 mg BID dose groups, respectively, with consistent findings observed for other secondary efficacy endpoints (Table 1). At M30, observed differences for mMayo remission between tofacitinib 10 and 5 mg BID were generally greater in the subgroup with a baseline endoscopic subscore of 1 vs 0 and in the subgroup with vs without prior tumour necrosis factor inhibitor (TNFi) failure (Table 1). The percentage of pts who experienced loss of remission by M30, as estimated from Kaplan-Meier curves, was numerically higher in the 5 vs the 10 mg BID dose group (28.1%; 95% confidence interval [CI], 17.68–39.49 vs 21.7%; 95% CI, 12.21–32.95, respectively). Table 2 shows adverse events of special interest, by dose group. One death occurred due to fatal coronavirus disease 2019 pneumonia (10 mg BID). Conclusion These long-term data showed that most pts in stable remission on tofacitinib 10 mg BID maintenance therapy maintained mMayo score remission through M30 after dose reduction to 5 mg BID. Dose group difference in remission at M30 was consistent with that at M6.1 Differences in remission between the tofacitinib 5 and 10 mg BID groups were greater in subgroups with an endoscopic subscore of 1 vs 0 and in subgroups with vs without prior TNFi failure. Overall, safety findings were consistent with tofacitinib’s known safety profile; incidence of serious infections and herpes zoster was numerically higher in the tofacitinib 10 vs 5 mg BID group. 1. Vermeire S et al. J Crohns Colitis 2021;15:1130-41.

P753 Comparing long-term outcomes in inflammatory bowel disease patients between switch and non-switch cohorts: no differences in infliximab persistence or safety outcomes after 54 months

Abstract Background Switching patients with inflammatory bowel disease (IBD) from originator to biosimilar infliximab has been demonstrated to be non-inferior to continuing originator infliximab by several studies. However, data comparing treatment outcomes beyond 12-months across switch and non-switch cohorts specific to IBD patients remains comparatively underreported. Here we report long-term (&amp;gt;48 weeks) treatment outcomes of the SAME study, a large-scale parallel cohort study of Australian IBD patients who underwent non-medical switching from originator to biosimilar (CT-P13) (n=204) or continued originator infliximab (n=141). Methods The SAME study was a multi-centre, prospective parallel cohort non-inferiority study across seven Australian hospitals over 48 weeks undertaken between May 2017 to October 2019. The data that we present here is the long-term extension study, obtained retrospectively, which sought to compare long-term infliximab persistence beyond 48 weeks across switch and non-switch cohorts. This study included five of the original seven sites, three of which switched from originator to biosimilar infliximab. The primary outcome was infliximab persistence, defined as the proportion of originator or biosimilar infliximab treated patients from the original cohort at last follow up. Secondary outcomes included disease worsening requiring infliximab dose escalation or discontinuation, rates of adverse events, and development of antibodies to infliximab (&amp;gt;10ng/mL). Results Data from 263 (90 Ulcerative colitis and 173 Crohn’s disease) of 345 (76%) patients enrolled in the original SAME study were included with median follow up of 54.2 months (IQR 46.1-59.3). At last follow-up, 103 (65.5%) and 68 (64.2%) patients across the switch and non-switch cohorts (p = 0.8) remained on infliximab. There were no differences in the proportions that discontinued infliximab due to clinical or biochemical worsening of disease (21.7 v 23.6%, p = 0.72); required infliximab dose escalation (35.2 v 32.4%, p=0.8); developed antibodies to infliximab (5.3 vs 11.3%, p = 0.09) or experienced drug related adverse events (7.8 vs 8.3%, p = 0.8) between switch and non-switch cohorts, at last follow-up. Conclusion Long-term infliximab persistence was similar between switch and non-switch cohorts after a median of 54 months. This study represent one of the longest ‘real-world’ comparisons between switch and non-switch cohorts specific to IBD, and should provide reassurance to clinicians and patients alike.

301 Tapinarof cream 1% once daily for the treatment of extensive atopic dermatitis in adolescents and children: 4-week maximal use trial

Abstract There is a need for efficacious non-steroidal topical therapies for atopic dermatitis (AD) without restrictions relating to duration, the extent of use and application sites. Tapinarof (VTAMA®; Dermavant Sciences, Inc., USA) is a first-in-class, non-steroidal, topical, aryl hydrocarbon receptor agonist approved by the Food and Drug Administration for the treatment of plaque psoriasis in adults and under investigation for the treatment of psoriasis in children down to 2 years of age and for AD in adults and children down to 2 years of age. Tapinarof cream 1% once daily (QD) demonstrated significant efficacy vs. vehicle at 12 weeks and was well tolerated in adults and adolescents and with moderate to severe AD in a previously reported phase 2b trial (NCT02564055). To assess the safety, tolerability, pharmacokinetics (PK) and efficacy of tapinarof cream 1% QD in adolescents and children with extensive AD. In this open-label, maximal use, phase 2 trial (NCT05186805), adolescents and children, aged 2–17 years, with a clinical diagnosis of AD, a validated Investigator Global Assessment in Atopic Dermatitis™ (vIGA-AD™) score ≥3, and percent body surface area (%BSA) involvement of ≥ 25% for adolescents (12–17 years) or ≥35% for children (2–11 years) were assigned tapinarof cream 1% QD for 27 days. Primary endpoints include adverse events, mean investigator-assessed local tolerability scores (overall and sensitive areas) and PK parameters. Secondary endpoints include a change in vIGA-AD™, Eczema Area Severity Index (EASI), %BSA affected, and Peak Pruritus-Numeric Rating Scale scores at each trial visit. Eligible patients completing this trial have the option to enrol in an open-label, long-term extension trial (NCT05142774) to receive up to an additional 48 weeks of tapinarof cream 1% QD treatment. Overall, 36 patients aged 2–17 years were enrolled at nine sites in the US and Canada. Of these patients, equal proportions (33.3% [12/36]) were young children (2–6 years), children (7–11 years) and adolescents (12–17 years); 66.7% (24/36) were male and mean age (standard deviation [SD]) was 8.9 (4.9) years. At baseline, 77.8% (28/36) had a vIGA-AD™ score of 3 (moderate) and 8/36 (22.2%) had a score of 4 (severe). At baseline, the mean (SD) %BSA affected was 52.4% (19.1%), 42.0% (10.0%) and 33.9% (8.6%) for young children, children and adolescents, respectively. The overall mean (SD) EASI score was 23.8 (9.2). This trial will assess the safety, tolerability, PK and efficacy of tapinarof cream 1% QD in 36 adolescents and children down to 2 years of age with extensive, moderate-to-severe AD.

330 Safety of abrocitinib in 3582 patients with moderate-to-severe atopic dermatitis with over 900 patients exposed for almost 2 years

Abstract Abrocitinib is an oral Janus kinase 1 inhibitor with demonstrated efficacy in moderate-to-severe atopic dermatitis (AD). To evaluate the long-term safety profile of abrocitinib and provide relevant information for practitioners. Data from patients who received ≥1 dose of abrocitinib 200 mg or 100 mg in the JADE clinical program were pooled in two cohorts. The consistent-dose (CD) cohort comprised patients who received the same abrocitinib dose during the entire exposure time in parent phase 3 studies and/or the long-term extension study JADE EXTEND (NCT03422822). Parent studies were: JADE MONO-1 (NCT03349060), JADE MONO-2 (NCT03575871), JADE TEEN (NCT03796676), JADE COMPARE (NCT03720470), JADE DARE (NCT04345367; 200 mg only) and JADE REGIMEN (NCT03627767); in JADE REGIMEN only patients from the open-label run-in phase (200 mg only) who did not subsequently enter the randomized phase were included. Patients from the phase 2b study (NCT02780167) were also included in the CD cohort. The variable-dose (VD) cohort comprised patients who received different doses of abrocitinib (200 and 100 mg) throughout exposure time in the parent study JADE REGIMEN and in JADE EXTEND, if enrolled. Patients included in the VD cohort completed the open-label period of JADE REGIMEN (abrocitinib 200 mg only) and entered the randomized phase (abrocitinib 200 mg, abrocitinib 100 mg or placebo); some patients subsequently entered the JADE REGIMEN rescue phase (abrocitinib 200 mg) and/or JADE EXTEND (abrocitinib 200 or 100 mg). The data cutoff was 16 April 2021. A Cox regression analysis evaluated risk factors for specific events of interest. Data from 3582 patients (4313.4 patient-years) were analysed: 2784 patients in the CD cohort (abrocitinib 200 mg, n = 1761, abrocitinib 100 mg, n = 1023) and 798 patients in the VD cohort. Duration of exposure was ≥48 weeks in 43.6%, 66.9% and 86.1% and ≥96 weeks in 18.0%, 23.2% and 45.4% of patients in the CD 200 mg, 100 mg and VD cohorts, respectively. In the CD and VD cohorts, the median age was 30.0 and 29.0 years; 490 (17.6%) and 145 patients (18.2%) were adolescents, and 143 (5.1%) and 30 patients (3.8%) were aged ≥65 years at screening, respectively. Incidence rates (IRs; unique patients with events per 100 patient-years) of serious adverse events (SAEs), deaths, and treatment-emergent adverse events of special interest are shown in Table 1. IRs for SAEs were consistent across abrocitinib doses but were higher in patients aged ≥65 years vs. younger patients. Serious infections were the most frequent SAEs; frequency was not dose related. Herpes zoster (HZ), herpes simplex and pneumonia were the most frequent serious infections. From a Cox regression analysis, potential risk factors for treatment-emergent HZ included abrocitinib dose, baseline age, medical history of HZ, absolute lymphocyte count &amp;lt;1000 mm−3 prior to infection and region. Most (95%) adjudicated opportunistic HZ infections were cutaneous; 2 (5%) were extracutaneous (one serious disseminated varicella-zoster virus infection, one serious HZ meningitis). IRs for nonmelanoma skin cancer, malignancy (excluding nonmelanoma skin cancer), thrombocytopenia and lymphopenia were higher in patients aged ≥65 years vs. younger patients but were not dose related. Abrocitinib has an acceptable long-term safety profile in this large analysis of patients with moderate-to-severe AD. Risk of SAEs, HZ infection, malignancies and hematologic abnormalities increased with older age; HZ risk was also related to dose and prior HZ infection.

Long-term safety and efficacy of tezepelumab in people with severe, uncontrolled asthma (DESTINATION): a randomised, placebo-controlled extension study

Tezepelumab is a human monoclonal antibody that blocks thymic stromal lymphopoietin. The drug has been tested previously in the phase 3 NAVIGATOR (NCT03347279) and SOURCE (NCT03406078) studies, and was subsequently approved as a treatment for severe asthma. This extension study recruited from NAVIGATOR and SOURCE and aimed to evaluate the long-term safety and efficacy of tezepelumab in individuals with severe, uncontrolled asthma. DESTINATION was a phase 3, multicentre, randomised, double-blind, placebo-controlled, long-term extension study. The study was done across 182 sites (including hospitals, clinics, medical centres, clinical trial centres, and private practices) in 18 countries. Participants (aged 12-80 years) were required to have good treatment compliance in the parent study. Randomisation was stratified by the parent study and all participants were re-randomised. Those who were previously randomised to receive tezepelumab in either parent study continued treatment of subcutaneous tezepelumab (210 mg every 4 weeks); those who were previously randomised to receive placebo in either parent study were re-randomised 1:1 to receive either subcutaneous tezepelumab (210 mg every 4 weeks) or placebo (every 4 weeks) using a randomisation list prepared by a computerised system. Total treatment duration (including the parent studies) was 104 weeks for all groups. Participants, investigators, and site staff were masked to treatment assignment. The primary endpoints were exposure-adjusted incidence of adverse events and serious adverse events and the secondary endpoint was the annualised asthma exacerbation rate; these were assessed from week 0 of the parent studies to week 104 of DESTINATION in all participants who were randomised and who received at least one dose of tezepelumab or placebo in either of the parent studies. The trial is registered with ClinicalTrials.gov, NCT03706079, and is closed to new participants. Participants were recruited between Jan 7, 2019, and Oct 15, 2020. For individuals who initially received tezepelumab (n=528) in NAVIGATOR, incidence of adverse events over 104 weeks was 49·62 (95% CI 45·16 to 54·39) per 100 patient-years, compared with 62·66 (56·93 to 68·81) for those receiving placebo (n=531; difference -13·04, 95% CI -17·83 to -8·18). For serious adverse events, incidence was 7·85 (6·14 to 9·89) per 100 patient-years for individuals who initially received tezepelumab and 12·45 (9·97 to 15·35) for those who received placebo (difference -4·59, -7·69 to -1·65). In SOURCE, incidence of adverse events was 47·15 (36·06 to 60·56) per 100 patient-years for those who initially received tezepelumab (n=74) and 69·97 (54·54 to 88·40) for those who received placebo (n=76; difference -22·82, -34·77 to -10·01). For serious adverse events, incidence was 13·14 (7·65 to 21·04) per 100 patient-years for those who initially received tezepelumab and 17·99 (10·66 to 28·44) for those who received placebo (difference -4·85, -14·88 to 4·53). Tezepelumab reduced the annualised asthma exacerbation rate over 104 weeks compared with placebo. In participants initially from NAVIGATOR, the annualised asthma exacerbation rate ratio over 104 weeks was 0·42 (95% CI 0·35 to 0·51); in those initially from SOURCE, the ratio over 104 weeks was 0·61 (0·38 to 0·96). Tezepelumab treatment was well tolerated for up to 2 years and resulted in sustained, clinically meaningful reductions in asthma exacerbations in individuals with severe, uncontrolled asthma. These findings are consistent with previous randomised, placebo-controlled studies and show the long-term safety and sustained efficacy of tezepelumab in individuals with severe, uncontrolled asthma. AstraZeneca and Amgen.