Background

Ustekinumab (Stelara) is a human immunoglobulin G1 kappa monoclonal antibody, available as a pre-filled syringe of 90 mg/1 mL for subcutaneous injection at a unit price of $4,593 and as a single-use vial of 130 mg/26 mL solution for intravenous (IV) infusion at a unit price of $2,080. The current review of ustekinumab is for the treatment of adult patients with moderately to severely active Crohn’s disease (CD) who have had an inadequate response with, loss of response to, or intolerance to either conventional therapy, including corticosteroids or immunomodulators, or to one or more tumour necrosis factor-alpha (TNF) antagonists, or who are corticosteroid-dependent.¹

Ustekinumab was previously reviewed by CADTH Common Drug Review (CDR) for psoriatic arthritis² and psoriasis.³ For psoriatic arthritis, the CADTH Canadian Drug Expert Committee’s final recommendation was that ustekinumab not be reimbursed.² In psoriasis, the former Canadian Expert Drug Advisory Committee had recommended that ustekinumab be reimbursed, subject to criteria or conditions.³ CDR also reviewed infliximab, adalimumab, and, most recently, vedolizumab, for moderate-to-severe CD. All three were recommended for reimbursement, subject to criteria or conditions.^4–6

The manufacturer submitted a cost-utility analysis (CUA) comparing ustekinumab with infliximab (brand and biosimilar), adalimumab, vedolizumab, and conventional therapy (including corticosteroids or immunomodulators) for the treatment of moderately to severely active CD. The analysis was conducted from a Canadian public-payer perspective over a 25-year time horizon. Two target populations were included: patients with moderately to severely active CD who had experienced a failure with conventional therapy only (FCTO), and those who had experienced a failure with anti-TNF therapy. The analysis also included a mixed population of the two. The CUA evaluated a dosage regimen of 90 mg every eight weeks or every 12 weeks for ustekinumab as well as a regimen reflecting the blend of the two dosages. The model structure consisted of a decision tree to model the induction-treatment phase and a Markov (cohort) structure to model maintenance treatment for the remainder of the time horizon. Model transition probabilities for the induction and the maintenance phases were based on network meta-analyses (NMAs) and the IM-UNITI trial assessing ustekinumab. The manufacturer’s base-case analysis did not include a cost for the IV administration of ustekinumab in the induction phase.

Summary of Identified Limitations and Key Results

CDR identified several key limitations with the model submitted by the manufacturer: uncertainty with the model transition probabilities and the utility values used. The former was mainly due to significant limitations and uncertainty with the NMAs used to populate the model transition probabilities, and the latter was because of inconsistency in how the utility values from the publication used by the manufacturer were implemented in the CUA. Other limitations identified by CDR concerned the data used for assessing infliximab, the adjustment of long-term transition probabilities using real-world evidence which favoured ustekinumab, and the uncertainty of the analysis in the long-term extrapolation of clinical data.

In the revised base case, CDR varied the health-state utility values and excluded the effect of real-world evidence on the transition probabilities after one year in the model. CDR also corrected an error in calculating the weighted average quality-adjusted life-years (QALYs) for the ustekinumab mixed dosage (every eight weeks/every 12 weeks) treatment option. This error appeared to default to the every eight weeks dosage, excluding the every 12 weeks dosage, and resulting in an overestimate of the incremental cost-utility ratio (ICUR) results for the ustekinumab mixed dosage.

The CDR base case for ustekinumab when compared with conventional therapy in the population experiencing FCTO resulted in an ICUR of $115,474 per QALY gained and in the population experiencing failure of anti-TNF therapy, $131,297 per QALY gained. For the mixed population, ustekinumab resulted in an ICUR of $119,058 per QALY when compared with conventional therapy.

Among the available biologic therapies in patients experiencing FCTO, ustekinumab every 12 weeks was the most cost-effective, with an ICUR of $115,474 per QALY compared with conventional therapy, followed by ustekinumab mixed dosage every eight weeks/every 12 weeks, with an ICUR of $623,571 per QALY when compared with ustekinumab every 12 weeks, then finally by ustekinumab every eight weeks, with an ICUR of $658,533 per QALY compared with ustekinumab mixed dosage. Other biologics were either dominated or subjected to extended dominance. In the patients who had experienced a failure with anti-TNF therapy, the most cost-effective treatment was biosimilar infliximab, with an ICUR of $90,277 per QALY compared with conventional therapy, followed by ustekinumab every 12 weeks with an ICUR of $228,571 per QALY compared with biosimilar infliximab. The remaining ustekinumab regimens (every eight weeks and mixed dosage) resulted in ICURs of more than $1 million per QALY. Remaining biologic therapies (adalimumab, infliximab and vedolizumab) were also dominated or subjected to extended dominance.

The manufacturer provided correspondence to this report indicating that the drug costs for the induction dose of ustekinumab would be reimbursed by the manufacturer. Excluding the drug costs incurred from the induction dose appears to improve the ICUR for ustekinumab when compared with conventional therapy in a population experiencing FCTO, with an ICUR of $95,442 per QALY gained, and in a population experiencing failure with anti-TNF therapy, with an ICUR of $77,840 per QALY gained. For the mixed FCTO and anti-TNF population, ustekinumab resulted in an ICUR of $91,260 per QALY compared with conventional therapy.

Conclusions

The efficacy and safety of ustekinumab compared with conventional and other biologic therapy were based on an indirect comparison with noted limitations and heterogeneity across studies that raise uncertainty over the comparative efficacy and safety of ustekinumab in both the induction and maintenance phases. Other key limitations of the economic model pertain to the utility values included and the effects of real-world evidence on transition probabilities. In light of these limitations, CDR suggests that the ICUR for ustekinumab ranges from $115,474 to $189,403 per QALY when compared with conventional therapy, and from being dominant to $870,045 per QALY when compared with other biologic therapies.

At an induction dose of 6 mg/kg followed by 90 mg at week 8 and every eight weeks thereafter, the cost of ustekinumab in year 1 ($33,798) and subsequent years ($29,855) is higher than the cost of vedolizumab ($26,320 and $21,458, respectively), adalimumab ($23,099 and $20,019, respectively), and infliximab (brand: $31,602 and $25,765, respectively; biosimilar: $16,800 and $13,697, respectively). When ustekinumab is administered every 12 weeks in the maintenance phase, the costs for year 1 ($24,612) and subsequent years ($19,904) are lower than or comparable to the other biologics, with the exception of biosimilar infliximab.

If the drug costs associated with the induction dose for ustekinumab are reimbursed by the manufacturer, the ICURs for ustekinumab compared with conventional therapy and other biologic therapies tend to improve, as would be expected.