Biologics targeting interleukin (IL)-23 and the risk of cardiovascular events: Where are we?

Abstract

Florence Poizeau

Hafid Ait-oufella

Hervé Bachelez

In a recent phase III trial assessing the efficacy and safety of the IL-23p19 inhibitor mirikizumab compared to placebo and secukinumab among moderate-to-severe plaque psoriasis patients, six cerebro-cardiovascular events occurred in the mirikizumab arm within the first 16 weeks versus none in the other arms.¹ The sponsor, Eli Lilly, decided to stop the development of mirikizumab in this indication. This rare signal raises the complex issue of the impact of IL-23 blockade on the risk of major adverse cardiovascular events (MACEs).

IL-23, composed of p40 and p19 subunits, plays a pathogenic role in psoriasis by driving the differentiation of naïve T cells towards IL-17-producing T cells. Monoclonal antibodies targeting p40 (IL-12/23 inhibitors) include ustekinumab and briakinumab, while p19 inhibitors include guselkumab, risankizumab, tildrakizumab and mirikizumab.

Experimental studies investigating the contribution of IL-23 to atherosclerosis have yielded conflicting findings. According to some models, IL-23p19 has a mainly anti-atherogenic function through effects on the gut microbiota and the intestinal barrier function.² Other models have shown that IL-17 responses, which are driven by IL-23, promote collagen synthesis by vascular smooth muscle cells, and maintain plaque stability.³ However, no intrinsic effect of IL-23 has been reported on plaque vulnerability to date.

The cardiovascular events in the mirikizumab trial echo previous concerns with the IL-12/23p40 inhibitor briakinumab, the development of which was stopped in phase III. The controversy extended to ustekinumab following a meta-analysis of data from phases II/III clinical trials, with both agents exhibiting an increased risk of MACE, although randomized controlled trials (RCTs) lack power to study rare events.⁴ Recently, an additional safety signal was issued when post-marketing surveillance flagged the IL-23p19 inhibitor risankizumab because of a disproportionately large number of cerebrovascular events, although under-reporting limits significance.^{5, 6}

Several cohort studies have concluded that ustekinumab and IL-23p19 inhibitors are safe with respect to the MACE risk, but here again conflicting results were reported.⁷ We will discuss three major pitfalls of observational studies evaluating the cardiovascular safety of biologics: (i) the transient nature of the risk (also called a ‘trigger effect’), (ii) interaction (i.e. ‘effect modification’) and (iii) confounding by indication. First, estimating incidence rates or hazard ratios assumes that the MACE risk is constant over time during biological treatment, whilst immunocardiologists suggest that IL-17 and/or IL-23 blockade could be detrimental in the short-term but beneficial in the long-term.⁸ Therefore, mixing high-risk and low-risk periods could mask any causal association between MACEs and a biologic. Consequently, longitudinal cohort studies should be considered with extreme caution. Second, investigating the MACE risk under biologics across all patients can dilute and overlook a MACE risk specific to some patients, for example those with pre-existing unstable atherosclerosis. Third, comparing the MACE risk across biologics assumes that these biologics are prescribed to similar patients. Otherwise, confounding by indication exists. Controlling for this bias is challenging, because the rationale for choosing a given medication is rarely assessed. The self-controlled approach, which was developed to investigate transient drug effects and to overcome confounding by indication, was used to investigate ustekinumab. Results suggested a transient increase in the MACE risk in the first 6 months of treatment, only among patients with several cardiovascular risk factors.⁹

Despite several MACE signals with IL-12/23 and IL-23 inhibitors in RCTs, uncertainty and controversy remain: Is there a short-term or a long-term risk? If any exists, is the risk class-dependent? Would the risk be mitigated by statins?¹⁰ Providing accurate answers to these crucial questions will require the integration of mechanistic scenarios, cutting-edge methodology and population-based studies.

None.

F Poizeau reports reimbursement for attending medical meetings by Novartis, Janssen and UCB; receiving fees as a consultant for Bristol Myers Squibb and Novartis. H Ait-Oufella reports receiving fees as a speaker for Abbvie, Pfizer, Lilly, UCB, Takeda, Servier and Fresenius Kabi; being Chief Scientific Officer and Chief Marketing Officer for POLYGON Therapeutics. H Bachelez reports receiving fees as a consultant for Anaptysbio, Boehringer-Ingelheim, Bristol Myers Squibb, Celtrion Healthcare, Eli-Lilly, Innovaderm, Janssen and Sanofi; serving as a speaker for Boehringer-Ingelheim, Bristol Myers Squibb and UCB; serving as an advisory board member for Boehringer-Ingelheim, Dermavant, Janssen, UCB; receiving grant support from Bristol Myers Squibb and Boehringer-Ingelheim.