One-year Outcomes With Ustekinumab Therapy in Infliximab-refractory Paediatric Ulcerative Colitis

A Multicentre Prospective Study

Jasbir Dhaliwal; Hayley E. McKay; Colette Deslandres; Jennifer Debruyn; Eytan Wine; Ashley Wu; Hien Huynh; Nicholas Carman; Eileen Crowley; Peter C. Church; Thomas D. Walters; Amanda Ricciuto; Anne M. Griffiths


Aliment Pharmacol Ther. 2021;53(12):1300-1308. 

In This Article


In this cohort of children and adolescents with extensive and treatment-refractory UC, ustekinumab successfully induced and, importantly, maintained steroid-free clinical remission to 1 year in a significant proportion. These prospective data, combined with the durability of remission in ongoing longer term follow-up and the relative simplicity of the regimen (subcutaneous administration without concomitant immunomodulator), suggest that ustekinumab will be an important addition to the therapeutic armamentarium for paediatric UC.

In comparison to adult-onset disease, UC developing during childhood or adolescence is more often extensive and associated therefore with a greater likelihood of either chronically active steroid-dependent disease and/or episodes of acute severe colitis. As documented in two recent paediatric multicentre North American inception cohorts, UC activity was mild at the time of the first presentation and treated primarily with 5-aminosalicylic acid (5-ASA) in only one-third of children.[1,4] Among the others, who presented with moderate or severe disease activity requiring steroids, 5-ASA was effective in maintaining steroid-free clinical remission at 1 year in only 20%.[4] Anti-TNF therapy, particularly infliximab, has constituted a significant advance in the treatment of paediatric UC.[5,6] Optimisation of dosing based on TDM reduces pharmacokinetic failure, but the mechanistic failure persists in at least 30% of children,[6] highlighting the need for alternate pathway biologic or small molecule therapies in a substantial proportion of young patients.

Our cohort was comprised exclusively of children and adolescents with prior failure of at least infliximab. The 44% steroid-free clinical remission rate (defined by PUCAI) at week 52 reflects the outcome of all patients on intent to treat basis. Among adult patients with prior anti-TNF exposure in the UNIFI trial, steroid-free clinical remission (defined as total Mayo score ≤2 and no subscore >1) was observed in 40% of week 8 responders at 44 weeks following the start of maintenance (i.e. 52 weeks following IV dose).[7] Corresponding rates were 33% among week 8 responders with prior failure of both anti-TNF and vedolizumab, and 51% among bio-naïve week 8 responders.[7] In our study, although the likelihood of continuing to receive ustekinumab at week 52 was similar among children who previously failed by two vs one class(es) of biologics, the likelihood of achieving steroid-free clinical remission was significantly less among those with prior vedolizumab and infliximab failure. Early abandonment of ustekinumab even prior to the start of maintenance may have been a factor in this small cohort. Nevertheless, the observation parallels the UNIFI clinical trial data and a recent real-world multicentre GETAID study of 103 adult UC patients.[11] In a multivariable model both clinical severity defined by partial Mayo score >6 (OR 0.10 [95% CI: 0.01–0.90], P = 0.04) and prior failure with anti-TNF therapy and vedolizumab (OR 0.03 [95% CI 0.01–0.42], P = 0.01), were associated with a decreased likelihood of achieving steroid-free clinical remission at weeks 12–16.[11]

The goal of medical treatment of children with UC is at a minimum steroid-free clinical remission.[12] When 5-ASA, thiopurines and optimised infliximab therapy fail to achieve this target, options for children are limited. Until the very recent paediatric UC adalimumab approval, no other biologics were approved by regulatory agencies for patients under 18 years of age. In our clinical practice prior to February 2018 only vedolizumab could be obtained off-label, and only in the setting of prior anti-TNF failure. Network meta-analysis of individual placebo-controlled clinical trial data, suggests the superiority of ustekinumab over vedolizumab in adult UC patients previously exposed to anti-TNF.[13] Our observational study further supports the sequencing of ustekinumab prior to vedolizumab in children with prior anti-TNF failure. In an ongoing analysis of long-term efficacy beyond week 52, the observed sustained efficacy among all remitters in our study is encouraging. The UNIFI long-term extension study similarly demonstrated sustained efficacy from week 44 of maintenance dosing through to week 92.[14]

This study is the first to report outcomes with ustekinumab in paediatric UC. One prior retrospective study in paediatric IBD included nine patients with UC in a mixed but predominantly Crohn's disease cohort of 52 children and young adults.[15] Clinical trials among adults are leading to expansion of the therapeutic armamentarium for IBD, but delays between adult and paediatric drug approval lead to considerable off-label use, often without the knowledge of optimal dosing regimens for children.[16] The industry-sponsored UNISTAR trial (NCT02968108) examined the pharmacokinetics and short-term (16 week) efficacy of ustekinumab in a dose-ranging multicentre study of 44 children with moderate to severely active Crohn's disease.[17] Patients weighing ≥40 kg (n = 26) were randomised to receive intravenously either 390 mg or a low dose of 130 mg, whereas children weighing <40 kg (n = 18) received 9 mg/kg or a low dose of 3 mg/kg.[17] Maintenance doses given via subcutaneous injection q8 weekly from weeks 8 to 16 were 90 mg and 2 mg/kg for patients weighing, respectively, ≥40 and <40 kg.[17] The primary goal was to compare paediatric pharmacokinetic data, using the same in-house proprietary electrochemiluminescent immunoassay (ECLIA) method, with those observed in reference adult populations in the ustekinumab development program.[18] Not surprisingly, early serum ustekinumab concentrations following an intravenous dose of 390 mg in patients weighing ≥40 kg were comparable to those observed with weight-range-based dosing of roughly 6 mg/kg in adults. In comparison, however, early serum ustekinumab concentrations trended lower even with the 9 mg/kg induction dose in those weighing <40 kg.[17] Paediatric pharmacokinetic studies of weight-based dosing with other biologic agents have similarly demonstrated lower exposure among the smallest children.[19,20] The planned ustekinumab induction dose for the phase 3 trial of ustekinumab in paediatric Crohn's disease is 9 mg/kg for patients weighing <40 kg. The smaller children with UC in our cohort received higher per kg doses of ustekinumab intravenously than did those weighing at least 40 kg, but we did not have access to early (weeks 8 or 16) ELISA measurements of ustekinumab serum concentrations to add to the UNISTAR observations.

Data derived from adult IBD cohorts concerning target levels of ustekinumab have been conflicting. Optimal thresholds associated with clinical or endoscopic remission have been determined to be significantly influenced by differences in assays utilised.[21] A comparison of measurements made using commercially available assays, one human mobility shift assay (HMSA) and two different ELISAs, in 61 serum samples demonstrated an almost twofold increased difference in ustekinumab concentrations between the HMSA and both ELISAs.[21] By far the most comprehensive pharmacokinetic and pharmacodynamic data come from the UNIFI trials, where serum ustekinumab concentrations were measured by in-house proprietary ECLIA method.[18] Serum concentrations of 3.7 mg/mL or greater at week 8 were associated with response to induction, and steady-state trough concentrations of 1.3 μg/mL or greater with the maintenance of efficacy (AUC 0.64; 95% CI: 0.58–0.69, sensitivity 74%; specificity 48%).[18] Importantly, the drug reached steady state by the second maintenance dose and was not affected by prior biologic or concomitant immunomodulator therapy.[18] A higher serum trough concentration of 4.5 μg/mL measured using a drug-tolerant HMSA after 26 weeks of treatment was reported to be associated with improved biomarker and endoscopic response in a real-world Crohn's disease cohort treated with subcutaneous ustekinumab for both induction and maintenance.[22] An agreement was found in one analysis between the Janssen ECLIA assay used in the pivotal trials and one ELISA, but certainly results must be interpreted with attention to the method used.

Whereas the UNIFI trial compared as maintenance q8 weekly vs q12 weekly subcutaneous dosing among responders to induction, the interval between doses for all our patients was initially 8 weeks, with shortening to 4 or 6 weeks in more than half. Full analysis of results from the STARDUST (NCT03107793) Crohn's disease trial is not yet published, but of patients in the treat to target arm starting maintenance with q8 weekly dosing, 60% escalated to q4 weekly dosing based on the ongoing elevation of Crohn's disease activity index and C-reactive protein or faecal calprotectin.[23]

Associations of ustekinumab levels with clinical, biomarker and endoscopic outcomes have been reported in quartile analyses of real-world IBD cohorts, in long-term extension phase of the licensing trials, and at week 16 endoscopy in the STARDUST trial,[23] but the role of TDM with ustekinumab is overall less clear than with anti-TNFs. Ustekinumab levels measured by ELISA during established maintenance in our study varied, as would be expected, with an interval between doses. The very early empiric interval shortening precluded meaningful assessment of adequacy of q8 weekly dosing across the paediatric age range. Nevertheless, drug availability was clearly adequate as reflected by a median drug concentration during maintenance of 5.2 μg/mL. Indeed, we observed higher drug exposure in those with ongoing colitis vs remission reflective of a pharmacodynamic rather than a pharmacokinetic failure. No anti-drug antibodies were detected in our cohort, consistent with the literature that ustekinumab exhibits low immunogenicity, thereby obviating need for concomitant administration of an immunomodulator. In the UNIFI trial, 5.7% of the population had evidence of anti-drug antibodies of which 44% were transient and 72% non-neutralising.[18]

Our study, although multicentre is clearly limited by the small sample size. Moreover, ustekinumab dosing regimens, rather than being protocolised, were at the discretion of paediatric gastroenterologists at CIDsCANN sites. The interval between ustekinumab injections was commonly shortened early following only one or two subcutaneous injections and empirically, without prior measurement of drug concentration in serum. As part of clinical practice, colonoscopic reassessment was not routinely performed in children achieving clinical remission, particularly during months of restricted access because of the COVID19 pandemic. Nevertheless, this study is the first to make use of prospectively collected data to examine short- and long-term clinical, biomarker and endoscopic outcomes in a cohort of children with UC, and to provide data concerning maintenance trough levels following induction dosing implemented in industry-sponsored paediatric Crohn's disease trials. Clinical trial design required to license ustekinumab in paediatric UC is not yet confirmed. Outcomes with ustekinumab monotherapy in our small cohort of biologic-refractory children with extensive colitis are encouraging. Clinical trials in bio-naive paediatric patients are warranted, as is further study of ustekinumab exposure particularly in young children early following induction as well as during maintenance to confirm optimal dosing.