Systematic Review With Meta-analysis

Loss of Response and Requirement of Ustekinumab Dose Escalation in Inflammatory Bowel Diseases

Hongsheng Yang; Bingyang Li; Qin Guo; Jian Tang; Bo Peng; Ni Ding; Miao Li; Qingfang Yang; Zicheng Huang; Na Diao; Xia Zhu; Jun Deng; Huili Guo; Pinjin Hu; Kang Chao; Xiang Gao


Aliment Pharmacol Ther. 2022;55(7):764-777. 

In This Article


Two decades after the first anti-TNF agent for IBD, ustekinumab, a new cytokine interleukin 12/23 blocker, has emerged as a valuable therapeutic option for patients with CD or UC.[5,6] However, similar to other biologics, a major concern is a reduction in response to ustekinumab. To the best of our knowledge, the present study is the first systematic review and meta-analysis to estimate the annual risk of ustekinumab LOR and dose escalation among primary responders, assess the regained response to dose escalation and summarise the predictors for LOR in patients with CD or UC. Our study shows that the overall annual risk of LOR with ustekinumab, among primary responders with CD, was 21% per person-year. Similarly, the annual risk of dose escalation among primary responders with CD was 25% per person-year. The pooled regained response rate of dose escalation in CD patients with secondary LOR was 58%. However, for patients with UC, no studies reported data on LOR with ustekinumab among primary responders, and only one study with 284 induction responders was included, leading to an overall annual risk of dose escalation of 18% per person-year. Fifty-eight per cent of secondary non-responders with UC regained symptomatic remission after dose escalation. These data emphasise the necessity of disease monitoring and the potency of dose optimisation during ustekinumab maintenance therapy.

For all approved biologics, LOR is a common concern in clinical practice. The annual risk of LOR among CD primary responders was 13% per person-year for infliximab,[34] 20% for adalimumab,[35] 21% for certolizumab-pegol[36] and 47.9% for vedolizumab,[37] as reported in published systematic reviews with meta-analyses. Most CD patients, in the aforementioned studies, taking biologics except infliximab, were exposed to anti-TNF drugs, possibly reflecting a refractory disease phenotype. Previous anti-TNF exposure or failure predicted a higher risk for LOR to biologic agents.[3] Therefore, the risk of LOR to ustekinumab is higher with infliximab and similar to adalimumab and certolizumab-pegol. However, the lower risk of LOR to ustekinumab compared to vedolizumab remains unexplained. The indirect comparison should be interpreted with caution because of the differences between patient characteristics, study design and analytic methodology. Thus, comparative trials are necessary to study the efficacy of the available treatment options for IBD.

To date, there has been no consensus on the definition of LOR. Various definitions have been used in studies on ustekinumab. This hampered our ability to accurately estimate the magnitude of the LOR. In contrast, the need for dose escalation among primary responders is a relatively objective and reliable measure. The consistency between the annual risk of LOR and the annual risk of dose escalation (21% vs 25% per person-year respectively) among primary responders with CD may reflect the reliability of our results. Interestingly, Ehrenberg et al. assessed the prevalence of dose escalation for approved IBD therapies by using the medical claims data of 7028 commercially insured IBD patients, which showed the proportion of IBD patients with ustekinumab dose escalation was 22% in the first 12 months during maintenance therapy.[38] We found that the annual risk of ustekinumab dose escalation was 25% per person-year for CD and 18% for UC. Despite the different types of data and study designs, the results were broadly similar.

Over half of CD patients with LOR to ustekinumab regained response after dose escalation. Both reinduction and interval reduction appeared to be effective. It has been shown that serum concentrations of ustekinumab are proportional to their dose and are associated with initial treatment efficacy.[39] In maintenance studies, the median trough concentration was approximately three-fold higher in patients administered ustekinumab q8w than in that administered q12w. Trough concentrations of 0.8 μg/mL (or even up to 1.4 μg/mL) or greater were associated with higher rates of clinical remission.[39] Dose escalation significantly increased ustekinumab trough levels; higher baseline and post-treatment ustekinumab levels were associated with better efficacy of dose escalation.[40] This supports the use of dose escalation in CD patients who lose response to ustekinumab and have low trough levels. Therapeutic drug monitoring may help to select which patients are most likely to benefit from dose escalation.

In contrast to recent reports showing relatively high rates of antibodies to infliximab (2.9%–60.8%) and adalimumab (0.3%–35%),[41] the rate of antibodies to ustekinumab through 1 year of treatment was only 2.3%,[5] indicating that ustekinumab has low immunogenicity. Such a low rate of anti-ustekinumab antibodies does not explain the relatively high annual risk of LOR in CD patients (21% per person-year). Although lower serum concentrations were observed among those who had antibodies to ustekinumab, there was no demonstrable effect of immunogenicity on its efficacy.[39] Moreover, Liefferinckx C, et al. have reported that antibodies to ustekinumab were not detected in ustekinumab-treated CD patients who lost their response, had an incomplete response or needed dose escalation.[40] These data suggest that immunogenicity may not be the main reason for LOR to ustekinumab, and it appears less valuable to monitor anti-ustekinumab antibodies to detect early LOR to ustekinumab in CD patients.

Several clinical features have been identified as predictors of LOR in patients with ustekinumab-treated CD. It appeared that ileal CD patients with a stricturing phenotype, highly active disease at induction/maintenance or previous biological therapy could have an increased risk for LOR. Concurrent immunomodulation therapy had a controversial role in predicting the risk for LOR. The value of predictors for LOR to ustekinumab should be interpreted with caution. These predictors were retrospectively analysed in a small cohort of patients and were not externally validated. More prospective studies with external cohorts are needed to investigate the predictors of LOR to ustekinumab in patients with CD. With substantial technological advances and increasing financial support for precision medicine, research encompassing multi-omic profiles, microbiomes and exposomes will further help to identify potential biomarkers, which may help clinicians to detect early LOR in clinical practice.

Very few studies have investigated the LOR or the need for dose escalation in patients with UC. The limited data showed that a relative proportion of these patients lost response to ustekinumab over time, and therapeutic benefits could also be obtained from dose escalation. Further studies are needed to provide more information on the LOR and the need for dose escalation in these populations.

Our study has several strengths including an extensive literature search and robust methodology. Of note, we used rigorous eligibility criteria and excluded studies that did not systematically document LOR or dose escalation in primary responders. However, there are some limitations to this study. First, the calculation of person-years in this study was approximated, and unintentional errors may have occurred. We have provided detailed and concise data on every individual rate of LOR or dose escalation according to treatment duration. Various methods of data presentation will help readers to better interpret the results. Second, the results were heterogeneous. Subgroup analyses were performed to explore heterogeneity. Subgroup analyses confirmed the stability of the results, but significant estimates of annual risk of dose escalation were detected based on induction regimen and study type. Moreover, a random effect model was used to analyse all outcomes prior to yielding a more conservative estimate. Third, definitions of LOR or dose escalation modality differed or were not reported within the included studies, which may, in part, explain the heterogeneity. Nevertheless, the subgroup analyses (by the definition of LOR and dose escalation modality) confirmed the stability of our results. Moreover, the annual risk of LOR and dose escalation was broadly similar, which reflected the reliability of our results.

In conclusion, our study shows that primary responders with CD experienced LOR to ustekinumab at a risk of 21% per person-year and required dose escalation at a risk of 25% per person-year. Fifty-eight per cent of secondary non-responders with CD may benefit from dose escalation. However, data on UC are scarce. Well-designed prospective studies are needed to investigate LOR and its predictors and biomarkers in patients with IBD, especially in patients with UC.