Ustekinumab in Paediatric Patients with Moderately to Severely Active Crohn’s Disease: Pharmacokinetics, Safety, and Efficacy Results from UniStar, a Phase 1 Study

Abstract Background and Aims The objective was to evaluate the pharmacokinetics, safety/tolerability, and efficacy of ustekinumab in children with moderately to severely active Crohn’s disease. Methods In this Phase 1, multicentre, 16-week, double-blind, induction dose-ranging study [NCT02968108], patients aged 2‐<18 years [body weight ≥10 kg] were randomised [1:1] to one of two weight range-based intravenous induction doses: 130 mg vs 390 mg in patients ≥40kg and 3 mg/kg vs 9 mg/kg in patients <40kg. At Week 8, all patients received a single subcutaneous ustekinumab maintenance dose of 90 mg in patients ≥40kg or 2 mg/kg in patients <40kg. Results A total of 44 patients were randomised and treated with ustekinumab [n = 23 lower dose; n = 21 higher dose]; median [interquartile range] age was 13.0 [12–16] years. Pharmacokinetics were similar to those in adults with Crohn’s disease. However, serum ustekinumab concentrations were lower among those with body weight <40 kg compared with patients ≥40 kg and the reference Phase 3 adult population. Through Week 16, 73% of patients reported ≥1 adverse event [82.6% lower vs 62% higher dose]; two discontinued due to adverse events [one in each group]. Serious adverse events occurred in 16% [26% lower, 5% higher dose], with Crohn’s disease exacerbation being the most frequent. At Week 16, 22%/29% [lower/higher dose] achieved clinical remission [Paediatric Crohn’s Disease Activity Index ≤10]. Conclusions The pharmacokinetics/safety profiles were generally consistent with those observed in adults with Crohn’s disease. These results suggest a different dosing regimen may be required for patients <40 kg from that employed in this study; additional pharmacokinetic analyses may be needed in this population.


Introduction
Childhood-onset Crohn's disease [CD] is most commonly ileocolonic; often evolves into complicated disease phenotypes; has high prevalence of extraintestinal manifestations; and, despite the use of immune-modifying therapies, often requires surgical resection. 1 In addition, children are more frequently exposed to steroids, which can lead to dependence. 2 Currently, there are only two approved biologic therapies [both tumour necrosis factor [TNF] antagonists] for the treatment of paediatric CD in North America and the European Union. As in adults, a sizeable portion of children respond initially and then lose response or become intolerant to current treatments, including TNF-antagonist therapies. 3 Thus, there is a considerable unmet medical need for novel therapeutic options that are safe, effective, and convenient for children with CD.
Ustekinumab is a first-in-class fully human immunoglobulin G1κ monoclonal antibody that binds with specificity to the p40 protein subunit of interleukins [IL]-12 and IL-23. These cytokines are involved in inflammatory and immune responses such as natural killer cell activation and CD4+ T cell differentiation and activation. 4 Ustekinumab is indicated for the treatment of moderately to severely active CD, ulcerative colitis, and active psoriatic arthritis in adults, as well as moderate to severe plaque psoriasis in adults and children. 4 The efficacy and safety of ustekinumab in adults with CD have been previously established in the UNITI-1 and UNITI-2 induction studies and in the IM-UNITI maintenance study. 5 Longterm ustekinumab maintenance therapy has been evaluated in the IM-UNITI long-term extension studies through 2 and 3 years. 6,7 In this Phase 1 study, we evaluated the pharmacokinetics [PK], safety/tolerability, and efficacy of ustekinumab in paediatric patients with moderately to severely active CD. Results through Week 16 are presented here; the long-term extension [LTE] is ongoing.

Study design
UniStar is a randomised, double-blind, dose-ranging study of intravenous [IV] ustekinumab induction followed by subcutaneous [SC] ustekinumab maintenance in paediatric patients with moderately to severely active CD [ Figure 1]. Within the 2 weeks preceding Week 0, patients underwent an ileocolonoscopy, and at Week 0, patients were randomised 1:1 into one of two treatment groups stratified by weight [<40 kg or ≥40 kg] and previous TNF-antagonist exposure status.

Dosing
The dosing regimens evaluated in this study were modelled to deliver ustekinumab exposure comparable to that observed in the corresponding reference adult CD population (ie, UNITI-1 [NCT01369329], UNITI-2 [NCT01369342], and IM-UNITI [NCT01369355] studies). 5,8 Patients in each treatment group received a single IV induction dose of ustekinumab at Week 0 as follows: • lower IV induction dose [herein referred to as 'lower dose'] of 3 mg/kg for patients 10

Patient population
Eligible patients were 2 to <18 years old in the USA, 6 to <18 years old elsewhere, of either sex, with a body weight of ≥10 kg and a diagnosis of moderately to severely active CD or fistulising CD for ≥3 months, with active colitis, ileitis, or ileocolitis confirmed at any time in the past by radiography, histology, and/or endoscopy. . Patients were required to be stable for at least 2 weeks on steroids before study start and stable for at least 4 weeks on immunomodulators before study start. Those who failed or were intolerant to anti-TNF therapy were eligible for participation. However, patients with exposure to anti-TNF biologic agents within 8 weeks of baseline and vedolizumab within 16 weeks of baseline were not eligible for entry into the study. Steroid tapering was left to the discretion of the investigator and could begin at Week 3 to allow more flexibility in the protocol. The protocol did specify that enrolled patients should not initiate or increase the dose of any of the following concomitant CD-specific medical therapies: oral 5-aminosalicylates [ASA], oral corticosteroids, immunomodulators [i.e. AZA, 6-MP, or MTX], and/or total parental or enteral nutrition up to Week 16. Patients who did initiate or increase the dose of these specific medical therapies up to Week 16 were considered treatment failures.

Outcomes
Ustekinumab serum concentrations and anti-drug antibodies [ADAs] were assessed. Serum ustekinumab concentrations and ADAs were evaluated using a validated drug-tolerant electrochemiluminescent immunoassay method on the Meso Scale Discovery [MSD®] platform [Gaithersburg, MD]. ADAs were detected in up to 100 μg/ mL of ustekinumab without interference. Safety was assessed by summarising the frequency and type of AEs.
The PCDAI is a validated multi-item measure of disease activity. 9 Clinical response [reduction from baseline in PCDAI score of ≥15 points] and remission [defined as a PCDAI score ≤10 points] at Week 16 was assessed. Daily diaries one week preceding scheduled visits were collected for abdominal pain, stool pattern, and general well-being [instead of 1-week history recall]. Diaries from the date of bowel preparation to the date of the procedure and the day after the procedure were excluded in the PCDAI calculation. The worst item [defined by the highest sub-score] across 7 days of non-excluded diaries for each sub-component were used for that visit. The patient had to complete four or more non-excluded evening diaries to calculate the history components of the PCDAI score.
Ileocolonoscopies were conducted at baseline and Week 16. Endoscopic response [reduction in SES-CD ≥50%] and remission [SES-CD ≤2] were assessed. To evaluate the SES-CD endpoints, patients with evaluable endoscopy data were considered eligible for the SES-CD analyses if they had a baseline SES-CD ≥3, excluding the contribution of the narrowing component score. These exclusion criteria were designed to ensure that there was active disease reached by endoscopy, not just a high score due to stricture.
A single reader at a central facility evaluated and scored all video endoscopies. The reader was blinded to treatment group throughout the duration of the study. Patients with at least one segment evaluated were considered to have evaluable endoscopy data. The reader recorded whether the endoscopy was evaluable or not evaluable, due to poor bowel preparation or other technical reasons. For each evaluable endoscopy, the reader scored each ileocolonic segment using the SES-CD and recorded the presence or absence of mucosal ulceration for each segment.

Statistical analysis
A sample size of 40 patients was chosen empirically for PK assessments based on experience from previous PK studies of other biologics in paediatric patients, including the study of golimumab in paediatric patients with ulcerative colitis. 10 Given the moderate variability associated with ustekinumab PK parameters in adult patients with CD and assuming comparable PK variability between adult and paediatric patients, a sample size of 40 was considered sufficient to provide adequate PK data in paediatric patients with CD. UniStar was considered to have met its primary objective if the PK of ustekinumab in paediatric patients with CD was demonstrated to be consistent with that previously observed in the reference adult CD population.
No formal hypothesis testing was conducted. Descriptive statistics [e.g., mean, median, standard deviation, interquartile range, minimum, and maximum] were used to summarise continuous variables. Counts and percentages were used to summarise categorical variables.
Patients who had any of the following events were considered to be treatment failures from the time of event onward: [1] a CD-related surgery thought to be a result of lack of efficacy of study agent [with the exception of minor procedures such as drainage of a superficial abscess or seton placement]; [2] discontinued study agent due to an adverse event [AE] of worsening CD or due to lack of efficacy; or [3] specified changes in concomitant CD medications.
Treatment failure rules were applied for all efficacy endpoints unless otherwise specified. In such a case, baseline values [at Week 0] were assigned from the point of treatment failure onward through Week 16, regardless of the observed data, for continuous endpoints and patients were considered as not achieving the respective endpoints for dichotomous endpoints. Treatment failure rules overrode other data handling rules.
For all analyses, patients with insufficient data for binary endpoints were considered to not have achieved their respective endpoint; for patients with insufficient data for continuous endpoints, the last available value was carried forward. These data handling rules were applied for all efficacy endpoints.
The total PCDAI score was calculated when three or more of the five components were available in full [i.e., all sub-components were complete] from the visit where the PCDAI was measured. If less than three components were completely available, then the PCDAI was not calculated based on the information collected at the visit. Conditional on three or more of the five component scores being available from the visit where the PCDAI was to be measured, in general, when the value of a sub-component within each of the five components was missing, the closest previous sub-component value was used. If all the sub-component scores of a component were missing, the closest previous component value was used.  adult Phase 3 CD studies [NCT01369329, NCT01369342, and NCT01369355; Figure 3]. At Week 0 following infusion, Week 3, and Week 6, respectively, mean serum ustekinumab concentrations were 51.3 μg/mL, 7.7 μg/mL, and 3.0 μg/mL following the low induction dose in the overall paediatric population, compared with 41.9 μg/mL, 8.4 μg/ mL, and 3.8 μg/mL following the 130-mg dose in adults. At the same time points, mean ustekinumab concentrations were 148.7 μg/ mL, 23.7 μg/mL, and 9.1 μg/mL, respectively, following the high induction dose in the overall paediatric population, compared with 125.2 μg/mL, 25.2 μg/mL, and 11.1 μg/mL following the ~6-mg/kg dose in adults.

Results
In the higher dose group, a pattern toward lower mean SUC was observed in patients weighing <40 kg versus those weighing ≥40 kg [ Figure 4]. Additional subgroup analyses were conducted to evaluate consistency of efficacy outcomes [clinical response and clinical remission at Week 8] by different baseline characteristics [demographics, baseline disease characteristics, and baseline concomitant medication]. Results were generally similar between subgroups and consistent with the overall study [data not shown].

Clinical and endoscopic outcomes
Overall, patients had endoscopies at baseline and Week 16 [n = 37] where 32% and 28% of patients had an endoscopic response in the lower and higher dose groups and 16% [n = 3] and 11.1% [n = 2] had achieved endoscopic remission, respectively [ Figure 5c].

Corticosteroid use
The mean steroid dose at baseline was 14. 4

Clinical benefit and pharmacokinetics
At Week 8, a greater proportion of patients were in clinical response in the higher ustekinumab concentration group compared with the lower ustekinumab concentration group [ Figure 7]. Similar trends were observed for median improvement from baseline in the PCDAI score. However, no correlation was observed between SUC and clinical remission at Week 8.

Safety/tolerability
Through Week 16, 73% of patients reported one or more AEs [  Score for Crohn's Disease; low dose, 3 mg/kg IV for patients <40 kg BW or 130 mg IV for patients ≥40 kg BW; high dose, 9 mg/kg IV for patients <40 kg BW or 390 mg IV for patients ≥40 kg BW. a Patients who had a prohibited Crohn's disease-related surgery, or who discontinued study agent due to an adverse event of worsening Crohn's disease or due to lack of efficacy or had prohibited concomitant medication changes, were considered not to be in endoscopic response or remission. In addition, patients with missing segments at the designated analysis time point had their baseline score for the missing segment[s] carried forward. Lower dose = 3 mg/kg IV for patients <40 kg BW or 130 mg IV for patients ≥40 kg BW; higher dose = 9 mg/kg IV for patients <40 kg BW or 390 mg IV for patients ≥ 40kg BW. group. A patient had pyrexia with a temperature of 38.1°C within 1 h of an infusion. The administration was stopped, and the patient's temperature returned to normal after a few minutes, so the administration was resumed.

Discussion
We conducted a study evaluating a range of ustekinumab doses in paediatric patients with moderate to severe CD. The primary objective of this study was met. Results show that serum concentrations of ustekinumab over time in the paediatric CD population were generally comparable to those in the reference adult CD population from the Phase 3 ustekinumab development programme [ Figure  3]. However, the subgroup analysis showed that SUCs were lower among children with body weight <40 kg compared with adolescents with body weight ≥40 kg and the reference Phase 3 adult population. This suggests that children with body weight <40 kg may require a different dose regimen from that employed in this study to attain similar exposures of ustekinumab compared with the typical adult population receiving the 390 mg induction dose.
With respect to immunogenicity, using a drug-tolerant assay, no patients had detectable antibodies to ustekinumab through Week 16, a finding consistent with the low incidence of immunogenicity in the reference adult CD population. Overall, results suggest that the PK and immunogenicity profiles are generally similar between paediatric and adult patients with CD and suggest that the higher dose induction regimen evaluated in this study, intended to provide ustekinumab exposures in paediatric patients commensurate with exposures observed in adult patients with CD, is appropriate in children ≥40 kg.
At Week 8, ustekinumab induction treatment resulted in meaningful improvements in the efficacy endpoints evaluated, including PCDAI-based clinical response and clinical remission. Although modest, overall improvements in objective measures of inflammation [reductions from baseline in CRP and faecal calprotectin] were generally reflective of improvements in efficacy outcomes. Overall, improvements in efficacy outcomes were considered generally similar between the two treatment groups and were observed as early as Week 3. Importantly, whereas definitive conclusions on the two dose regimens cannot be made based on efficacy results alone, the overall benefit profile observed was consistent with benefits observed in the adult CD studies.
Baseline disease characteristics were representative of a paediatric population of patients with moderately to severely active CD, with the majority having failed one or more anti-TNF agents. Overall, demographic and baseline characteristics including median duration of disease and median PCDAI score, a major efficacy variable, were generally well balanced across the two treatment groups. Although the proportions of patients with abnormal levels of inflammatory markers at baseline were similar across treatment groups, overall baseline values for these and SES-CD measures were noted to be somewhat greater for patients in the lower than in the higher induction dose group. Notably, whereas some heterogeneity, largely attributable to the limited sample size, was observed in baseline values, overall inflammatory burden suggested by CRP and faecal calprotectin was generally consistent with that observed in the ustekinumab adult CD programme.
Ustekinumab induction and short-term maintenance treatment through Week 16 were generally well tolerated in this paediatric CD population. Although SAEs and CD events were higher in the lower dose group, there was no evidence of an increased risk of infections in either group. But again, the rate of CD-related AEs in the lower dose group is notable. There were no injection site reactions, and one AE temporally related to infusion was reported in the lower induction dose group. No anaphylactic or serum sickness-like reactions were reported. Although interpretation is limited by small sample sizes, no patterns  were observed suggesting a different safety profile in the younger or smaller [<40 kg] participants compared with the older participants [aged 12 to 17 years] or those ≥40 kg. Overall, safety and tolerability results from this study suggest that the safety profile of ustekinumab in the paediatric CD population was generally consistent with the established safety profile of ustekinumab in the adult CD population.
In conclusion, improvements in clinical and endoscopic disease activity were observed across the efficacy outcome measures evaluated. Reductions were seen in objective biomarkers of inflammation in this treatment-refractory group of children with CD. The pharmacokinetics and safety profiles were consistent with those reported for adults with CD treated with ustekinumab. Overall, our data favour    AE, adverse event; BW, body weight; CD, Crohn's disease; IV, intravenous; Na, not assessed; PCDAI, Paediatric Crohn's Disease Activity Index; UST, ustekinumab. a Including budesonide. b Patients who had a prohibited Crohn's disease-related surgery or discontinued study agent due to an AE of worsening CD, or due to lack of efficacy, or had prohibited concomitant medication changes, were considered not to be in clinical remission and not receiving corticosteroids. c Patients who had insufficient data to calculate the PCDAI score at that visit were considered not to be in clinical remission and not receiving corticosteroids. d Patients who had a missing value in corticosteroid use at Week 8 [Week 16] had their last value carried forward. e Low dose, 3 mg/kg IV for patients <40 kg BW or 130 mg IV for patients ≥40 kg BW. f High dose , 9 mg/kg IV for patients <40 kg BW or 390 mg IV for patients ≥40 kg BW.