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


Twenty-five children with UC (28% male; median age at ustekinumab start 14.8 years [IQR 12.3–16.2]) received at least one dose of ustekinumab following the failure of prior biologic therapy at one of six CIDsCANN sites. The median duration of colitis was 2.3 years (IQR 1.1–4.2) at the time of ustekinumab initiation. Table 1 describes patient characteristics. All patients had failed prior anti-TNFα (all infliximab, 3 also adalimumab) and 12 (48%), additionally, vedolizumab therapy. Two-thirds of children (n = 17) were receiving concomitant corticosteroids at ustekinumab initiation.

Treatment Regimens

All patients received ustekinumab via intravenous induction, 260 mg in 13 weighing 18.3–52.6 kg; 390 mg in 11 weighing 36–76.1 kg and 520 mg in 1 weighing 74.8 kg. The median (IQR) ustekinumab dose was 6.4 (5.5–7.5) mg/kg. Children weighing <40 kg (n = 8) received a higher per kg dose (median 7.7, IQR 6.7–9.7), than those weighing ≥40 kg (n = 17) (median 5.7, IQR 5.4–7.0), P = 0.002. All patients received ustekinumab without a concomitant immunomodulator. Subcutaneous maintenance injections were scheduled to begin at week 8, using 90 mg for all except the smallest child, who received 45 mg. The standard q8 weekly dosing interval could be shortened after the first subcutaneous dose at the discretion of the treating physician based on continuing symptoms.

Clinical Outcomes

Figure 1 displays patient disposition. On an intent-to-treat basis, 11 (44%) of 25 met the primary endpoint of week 52 steroid-free clinical remission while receiving ustekinumab. This included 9 (69%) of 13 who previously treated with anti-TNF only vs 2 (17%) of 12 who previously failed also by vedolizumab, P = 0.008. Rates of steroid-free clinical remission at 6 and 12 months are displayed in Figure 2. The increase in the proportion of patients in steroid-free clinical remission at 6 vs 12 months was due to final tapering of low-dose prednisone after 6 months (n = 2 patients), or persistence at 6 months of mild symptoms, which subsequently resolved (n = 3). In a multivariable model, the odds ratio of achieving steroid-free clinical remission at week 52 following prior failure of only anti-TNF therapy vs additionally vedolizumab was 17.4 (95% CI 1.6–190), adjusted for age and sex. Of the 11 patients achieving the primary endpoint at 52 weeks, 8 initiated ustekinumab early enough to allow more than 1-year follow-up. Among these eight, steroid-free clinical remission has to date been further maintained to ≥72 weeks (maximum 147 weeks) in seven, including two children with prior anti-TNF and vedolizumab failure, and five with prior anti-TNF failure only (see Figure 3).

Figure 1.

Disposition of children and adolescents with refractory ulcerative colitis commencing ustekinumab

Figure 2.

Rates of steroid-free clinical remission (SFCR) at weeks 26 and 52 and sustained with no steroid use during the interval (26–52 week). Intent-to-treat analysis following ustekinumab induction among cohort overall, and comparing those with prior failure of only one class of biologics (anti-TNF) vs two classes (anti-TNF and vedolizumab). Endoscopic improvement in patients achieving SFCR was based on colonoscopic reassessment, or its surrogate faecal calprotectin <250 μg/g, if repeat colonoscopy not performed

Figure 3.

Likelihood of continuing to receive ustekinumab sole biologic therapy: Kaplan-Meier curves stratified by prior biologic exposure

Figure 3 graphs the likelihood of continuing ustekinumab to 1 year and beyond. At week 52, overall, 16 (64%) of 25 patients continued to receive ustekinumab therapy, including 9 (69%) of 13 with only prior anti-TNF failure, and 7 (58%) of 12 with prior failure also of vedolizumab (P = 0.57). Corticosteroids were discontinued in all after median 19.4 (IQR 5.5–22.5) weeks. The five non-remitters continuing ustekinumab (all with prior failure of both infliximab and vedolizumab) were off steroids but with mild to moderate disease activity (median PUCAI 20, IQR 12.5–35). Four of these five had improved from baseline with the median difference in PUCAI of −25 (−53 to 0).

Five patients abandoned ustekinumab after receiving one intravenous dose, including four (three with prior failure of both vedolizumab and infliximab) who underwent colectomy due to ongoing active colitis. One (previously treated with infliximab only) stopped therapy by family request. Another four patients discontinued ustekinumab during the maintenance phase of therapy, one initiating tofacitinib, one vedolizumab and two undergoing colectomy. The median time to discontinuation of ustekinumab was 17.6 weeks (IQR 1.6–31.8) among children with only anti-TNF failure vs 0 weeks (IQR 0–26.6) among those with prior failure of two classes of biologics, P = 0.71 (Figure 3). In total 6 (24%) of the cohort underwent colectomy by 1 year at a median time of 5.3 weeks (IQR 3.6–36.4) following intravenous ustekinumab.

Among patients initiating subcutaneous ustekinumab, dosing interval was shortened during maintenance in 15 (75%) of 20 (to q4 weekly in 12 and q6 weekly in three), with the majority of adjustments made early following the first or second subcutaneous dose (8 [53%] and 5 [33%] respectively) and empirically based on continuing symptoms. Low ustekinumab concentration (0.958 μg/mL) was the reason for interval shortening in only one patient, who was nevertheless already in steroid-free clinical remission. All nine with prior failure of two classes of biologics had interval shortened, compared to 6 of 11 with prior infliximab failure only, P = 0.004.

Ustekinumab Concentrations and Antidrug Antibodies

Fourteen (70%) of 20 patients initiating maintenance therapy had serum ustekinumab levels and anti-drug antibody levels measured via ELISA, at the median time of 30.9 weeks (IQR 22.1–55.8). The median ustekinumab trough concentration was 5.2 μg/mL (IQR 4.0–9.7) and no patients developed detectable antibody levels. The median trough levels were higher with q4 vs q8 weekly dosing, respectively 9.1 μg/mL (range 4.4–12.7) vs 4.2 μg/mL (range 3.5–5.4), P = 0.08, but greater exposure was not associated with a superior rate of clinical remission. During the maintenance phase of therapy median ustekinumab levels were higher when measured in patients with clinically active colitis (all with dosing interval already empirically shortened to 4 weeks) than in patients in clinical remission (respectively, 9.5 μg/mL [IQR 4.7–11.8] vs 3.9 μg/mL [IQR 2.7–6.2]), P = 0.02.

Endoscopic Outcomes and Biomarker Remission

To verify more than symptom resolution, the 11 patients meeting the primary endpoint of steroid-free clinical remission at week 52 underwent colonoscopic reassessment (n = 5) and/or faecal calprotectin determination (n = 9). Seven had evidence of endoscopic improvement defined by Mayo endoscopic subscore of 0 or 1, or by the surrogate of FCP <250 μg/g if colonoscopic re-evaluation was not performed. FCP was <250 μg/g in five of nine remitters with stools examined at the median time of 51.1 weeks (IQR 34.6–56.0), representing a significant decline from baseline (median FCP 863, IQR 759–2100). Mayo endoscopic score was 0 or 1 in four of the five clinical remitters undergoing colonoscopy at the median time of 51.4 weeks (IQR 29.1–91.5), all whose baseline Mayo endoscopic score had been ≥2.


No adverse events were reported during ustekinumab therapy.