TA888 · STA

Risankizumab for previously treated moderately to severely active Crohn's disease

Recommended with restrictionsApril 2023

Only recommended if the disease has not responded well enough or lost response to a previous biological treatment, or a previous biological treatment was not tolerated, or TNF-alpha inhibitors are not suitable. Only available after CE mark for on-body device is granted. Commercial arrangement (simple discount patient access scheme) required.

Source documents

Intervention

risankizumab (Skyrizi)
Biological DMARD · IL-23 modulator · subcutaneous injection

Conditions

moderately to severely active crohn's diseasegastroenterology · relapsed_refractory
crohn's diseasegastroenterology · relapsed_refractory

Comparators

NameType Established Committee preferred
adalimumabactive drugYes
infliximabactive drugYes
ustekinumabactive drugYes
vedolizumabactive drugYesYes

Clinical trials

TrialDesignPhasePivotal
ADVANCERCT3Yes
MOTIVATERCT3Yes
FORTIFYRCT3Yes

Economic model

partitioned survival (company)
Time horizon: 10 years (cost-comparison model)

Methodological decisions (9)

equivalence assumption

Clinical equivalence of risankizumab versus comparators for cost-comparison analysis

Company: Network meta-analyses showed risankizumab similarly clinically effective to comparators

ERG: Network meta-analyses did not support equivalent clinical effectiveness; wide credible intervals and point estimates uncertain

Committee: Only enough certainty that risankizumab has at least equivalent benefits to vedolizumab; uncertainty regarding TNF-alpha inhibitors and ustekinumab

ICER impact: uncertain_direction

indirect comparison method

Company proposed splitting network meta-analyses into 2 separate networks (risankizumab and ustekinumab vs adalimumab, infliximab and vedolizumab) based on drug mechanism of action and other characteristics. EAG and committee preferred a single network approach, noting that network connections should be based on comparator connections rather than drug characteristics.

Company: Split networks approach because of differences in drug mechanism of action, induction duration and half-life, and because single network analyses lacked face validity with placebo remission rates appearing too high.

ERG: Single network more appropriate; connections should be based on comparator connections not drug characteristics. Ustekinumab and vedolizumab have similar half-life and are comparable treatment options.

Committee: Single network approach

ICER impact: uncertain_direction

indirect comparison method

Company used fixed effects model for network meta-analyses. EAG and committee preferred random effects model given differences between trials in baseline risks, stratification, and observed temporal effect in placebo remission rates.

Company: Fixed effects model; exploration of random effects model produced wide confidence intervals and values favouring placebo over biological treatments.

ERG: Random effects model more appropriate given observed differences between trials including temporal effect on placebo remission rates.

Committee: Random effects model

ICER impact: uncertain_direction

indirect comparison method

Company's risk difference approach without adjustment for heterogeneity or temporal effect. EAG preferred adjustment for temporal effect observed in placebo remission rates across trials.

Company: Risk difference approach without explicit adjustment for temporal effect or baseline risks.

ERG: Preferred to include adjustment for temporal effect observed in placebo remission rates.

Committee: Adjustment for temporal effect needed

ICER impact: uncertain_direction

indirect comparison method

Committee noted at first meeting that updated analyses were needed using risk ratios rather than risk differences to compare risankizumab with other biological treatments, as this may be more informative given study heterogeneity and would allow more straightforward exploration of data to improve precision.

Company: Not explicitly stated for risk ratio approach in this chunk

ERG: Not explicitly stated for risk ratio approach in this chunk

Committee: Risk ratios rather than risk differences, exploring both random effects and fixed effects models

ICER impact: uncertain_direction

indirect comparison method

Network meta-analysis approach: fixed vs random effects model and risk difference vs risk ratio

Company: Risk difference approach with wide confidence intervals; company explored random effects but noted results favoured placebo

ERG: Random effects model with risk ratios rather than risk differences; adjustment for temporal effect in placebo remission rates

Committee: EAG approach preferred - random effects model using risk ratios with adjustment for temporal effect, exploring both random and fixed effects

ICER impact: uncertain_direction

model structure

Cost-utility model structure based on CDAI health states vs reflective of clinical pathways

Company: CDAI-based model with decision tree induction phase and Markov maintenance phase

ERG: Model does not reflect lifelong relapsing-remitting nature and current treatment pathways

Committee: Model not suitable for decision-making due to not reflecting treatment pathway with multiple biological treatments

ICER impact: increases

stopping rule

Maximum treatment duration assumption in cost-utility model

Company: 1-year maximum treatment duration, consistent with previous NICE appraisals for Crohn's disease

ERG: 20-year maximum treatment duration reflecting lifelong nature of Crohn's disease

Committee: 20-year maximum treatment duration more reflective of clinical practice

ICER impact: decreases

treatment sequencing

Whether model should allow multiple sequential biological treatments or only one

Company: Model assumes all people have conventional care after first biological treatment

ERG: Model should reflect lifelong relapsing-remitting nature allowing multiple biological treatments

Committee: Model must reflect current clinical pathway where people can have more than 1 biological treatment

ICER impact: increases

Evidence gaps

no direct comparisonLack of direct comparative evidence between risankizumab and comparators (adalimumab, infliximab, ustekinumab, vedolizumab), requiring network meta-analyses which have significant methodological challenges.
surrogate not validatedPrimary outcomes of clinical remission (CDAI) and endoscopic response (SES-CD) are measured in trials but CDAI is not used in clinical practice; Harvey-Bradshaw Index is used instead. While considered broadly comparable, this represents a difference between trial and clinical practice measures.
no direct comparisonNo clinical trials directly comparing risankizumab with other biological treatments; network meta-analysis based on indirect evidence with wide credible intervals
surrogate not validatedCDAI used in model to define clinical response and remission is not used in clinical practice; correlates with Harvey-Bradshaw Index but model structure limits applicability

Commercial arrangement

simple discount pas · confidential · critical for recommendation

Special considerations

Innovation acknowledged