TA882 · STA

Voclosporin with mycophenolate mofetil for treating lupus nephritis

Recommended with restrictionsMarch 2023

Only recommended if the company provides voclosporin with mycophenolate mofetil according to the commercial arrangement (simple discount patient access scheme)

Source documents

Intervention

voclosporin (Aurquivia)
Calcineurin inhibitor · calcineurin inhibition · oral

Conditions

lupus nephritisrheumatology · active
lupus nephritisrheumatology · chronic

Comparators

NameType Established Committee preferred
mycophenolate mofetil alonestandard of careYesYes
tacrolimus with mycophenolate mofetilactive drugYesYes
tacrolimusactive drugYes

Clinical trials

TrialDesignPhasePivotal
AURORA 1RCTPhase 3Yes
AURORA 2RCTPhase 3Yes

Economic model

markov (company)
Time horizon: approximately 72 years
Cycle length: unclear from excerpt

ICER

£20,000–£30,000 (voclosporin plus mycophenolate mofetil vs mycophenolate mofetil alone) · high uncertainty

Methodological decisions (12)

cost assumption

Additional monitoring costs for voclosporin versus tacrolimus

Company: voclosporin does not require additional monitoring because it has predictable pharmacokinetics allowing fixed dosing

ERG: inconsistent with costs modelled for tacrolimus

Committee: extra monitoring costs for voclosporin not appropriate; voclosporin does not need therapeutic drug monitoring like tacrolimus

ICER impact: negligible

cost assumption

Treatment discontinuation for non-trial comparators (tacrolimus, standard of care)

Company: initially assumed no discontinuation due to lack of evidence

ERG: clinically implausible; preferred to assume discontinuation equal to mycophenolate mofetil

Committee: company updated to assume discontinuation equal to voclosporin; assumptions did not have significant impact on cost-effectiveness

ICER impact: negligible

indirect comparison method

Fixed effects vs random effects network meta-analysis

Company: Fixed effects network meta-analysis justified because random effects estimates were not converging

ERG: Random effects network meta-analysis should be implemented using informative priors to account for considerable variation across trials

Committee: Random effects network meta-analysis was preferred and implemented in company's updated base case

ICER impact: negligible

indirect comparison method

Network meta-analysis heterogeneity across populations and trial outcomes

Company: network meta-analysis was appropriate

ERG: appropriate but heterogeneity increased uncertainty

Committee: heterogeneity noted; longer-term extrapolations particularly uncertain when associated with heterogeneity

ICER impact: increases

model structure

Whether people with stages 3b to 4 CKD could have a response

Company: People with stages 3b to 4 CKD modelled as unable to have response; provided scenario analyses showing 2.5% response rate but did not implement in base case

ERG: Clinical experts agreed response rate should not be zero and this should be reflected in base case

Committee: Assumption that people with stages 3b to 4 CKD could not have a response was acceptable; cost-effectiveness estimates insensitive to this assumption

ICER impact: negligible

model structure

Method for deriving transition probabilities from AURORA trials

Company: Used 'count method' due to limitations with alternative statistical methods which provided unrealistic outcomes

ERG: Acknowledged company's efforts but noted inherent uncertainties remain with count method

Committee: Company's explanation accepted; uncertainty remains

ICER impact: uncertain_direction

model structure

Company modelled people with stages 3b to 4 CKD as unable to have a response; EAG and clinical experts disagreed. Company provided scenario analyses allowing 2.5% response but did not implement in base case.

Company: no response for stages 3b to 4 CKD is reasonable; impact is small

ERG: clinical experts agreed number with response was not zero and should be reflected in base case

Committee: cost-effectiveness estimates were insensitive to this assumption; not a key uncertainty

ICER impact: negligible

other

Proportion of people with stage 5 CKD receiving kidney transplant

Company: Original base case used estimate considered too high

ERG: 65% transplantation rate based on clinical advice

Committee: 65% transplantation rate accepted

ICER impact: uncertain_direction

other

Corticosteroid doses used in AURORA trials compared to NHS clinical practice

Company: Lower doses used in trials are effective and in line with guidelines

ERG: Agreed steroid doses were lower than clinical practice but still effective

Committee: AURORA trials generalisable to UK but steroid doses may not reflect established NHS clinical practice

ICER impact: uncertain_direction

other

39.5% of AURORA 1 participants did not enrol in AURORA 2; impact on discontinuation analysis

Company: Provided scenario analyses with different assumptions for discontinuation (last observation carried forward, and two alternative scenarios)

ERG: Only last observation carried forward scenario was appropriate

Committee: Company efforts satisfactory; last observation carried forward scenario likely clinically plausible and slightly reduces cost-effectiveness

ICER impact: decreases

stopping rule

Treatment duration modelled at 36 months in line with AURORA data

Company: clinical experts supported stopping at 36 months; provided scenarios for 12 and 18 month durations

ERG: scenarios helpful but uncertain because retreatment not included

Committee: 36 months assumption arbitrary; almost half of people stopped by this point; clinical experts suggested ~9-12 months expected; key uncertainty is that retreatment not included in model

ICER impact: increases

treatment effect waning

Long-term treatment effect extrapolation beyond 36 months of trial data

Company: assumed long-term transition probabilities equal average of month 30 and 36 data with treatment waning; for active disease and partial response, voclosporin effect equals mycophenolate mofetil alone after 36 months; for complete response, equals average of both

ERG: considerable uncertainty using short-term data to predict off-treatment long-term outcomes; preferred to assume voclosporin plus mycophenolate and mycophenolate alone equal based on average transition probabilities

Committee: longer-term efficacy difficult to establish from short-term data; uncertainty with both approaches; model flexibility prevented EAG implementing preferred base case

ICER impact: increases

Evidence gaps

no direct comparisonLack of direct evidence comparing voclosporin plus mycophenolate mofetil with other relevant comparators besides mycophenolate mofetil; required network meta-analysis
no uk dataAURORA trials included no people from the UK; generalisability to NHS limited
short follow upClinical experts raised concerns about lack of long-term evidence for voclosporin; uncertain how it would be used in clinical practice
short follow upAURORA 1 and AURORA 2 data on voclosporin plus mycophenolate mofetil covered 3 years; extrapolation needed for approximately 69 further years
no direct comparisonNo evidence comparing voclosporin with higher and lower doses of steroids
immature overall survivalNo long-term efficacy data; stakeholders noted long-term treatment effect is unproven assumption

Commercial arrangement

simple discount pas · confidential · critical for recommendation

Special considerations

Equality issues raised