TA877 · STA

Finerenone for treating chronic kidney disease in people with type 2 diabetes

Recommended with restrictionsTechnology Appraisal Committee BFebruary 2023

Recommended as an add-on to optimised standard care (ACE inhibitors or ARBs at highest tolerated doses, and SGLT2 inhibitors) for stage 3 and 4 CKD with albuminuria associated with type 2 diabetes in adults with eGFR of 25 ml/min/1.73 m² or more.

Source documents

Intervention

finerenone (Kerendia)
non-steroidal mineralocorticoid receptor antagonist · mineralocorticoid receptor antagonist (non-steroidal) · oral

Conditions

type 2 diabetesmetabolic_endocrine · chronic
chronic kidney diseaserenal · chronic
chronic kidney disease (stage 3 and 4 with albuminuria) associated with type 2 diabetesrenal · chronic
chronic kidney disease associated with type 2 diabetesrenal · CKD stage 3 and 4 with albuminuria
chronic kidney disease associated with type 2 diabetesmetabolic_endocrine · chronic

Comparators

NameType Established Committee preferred
ace inhibitors (at maximum tolerated licensed dose)active drugYes
angiotensin-receptor blockers (arbs) (at maximum tolerated licensed dose)active drugYes
sglt2 inhibitorsactive drug
placebo (plus standard care)placebo
placebo plus standard careplacebo
standard care including sglt2 inhibitorsactive drugYes
standard care with ace inhibitors or arbsstandard of careYes
standard care with ace inhibitors or arbs and sglt2 inhibitorsstandard of careYes
placeboplacebo
standard care alonestandard of care
standard care with sglt2 inhibitorsstandard of care

Clinical trials

TrialDesignPhasePivotal
FIDELIO-DKDRCTphase 3Yes
FIGARO-DKDRCTphase 3Yes
FIDELITYmeta_analysisunknown

Economic model

markov (company)
Time horizon: 34.2 years lifetime
Cycle length: 4 months
markov
Time horizon: 34 years
Cycle length: annual

ICER

Below £20,000 (finerenone plus standard care vs standard care alone) · high uncertainty

Methodological decisions (14)

comparator selection

Whether SGLT2 inhibitors should be included as a relevant comparator for finerenone

Company: SGLT2 inhibitors should not be included as a comparator because they were not established NHS practice at time of company submission; company focused on finerenone as second-line treatment added to ACE inhibitors/ARBs without SGLT2 inhibitors

ERG: SGLT2 inhibitors should be considered a relevant comparator despite not being established at time of trial, because recent NICE recommendations will likely increase uptake

Committee: SGLT2 inhibitors are a relevant comparator, but comparison of finerenone with SGLT2 inhibitors in SGLT2 inhibitor-naive population is missing; finerenone can only be recommended as option in addition to SGLT2 inhibitors or when unsuitable

ICER impact: uncertain_direction

comparator selection

Whether to include FIGARO-DKD and FIDELITY data in addition to FIDELIO-DKD for decision making

Company: FIDELIO-DKD should be the primary evidence base; combining FIDELIO-DKD and FIGARO-DKD data was not pre-specified and questionable statistically

ERG: Company provided insufficient details for critique; additional evidence from FIGARO-DKD could reduce uncertainty around the marketing authorisation population

Committee: Further evidence from FIGARO-DKD and FIDELITY are relevant and appropriate; FIGARO-DKD data support FIDELIO-DKD results but have limitations

ICER impact: uncertain_direction

model structure

Whether to include clinical evidence from FIGARO-DKD and FIDELITY meta-analysis in addition to FIDELIO-DKD

Company: Company excluded FIGARO-DKD from initial evidence base because full data was not available; suggested combining FIDELIO-DKD and FIGARO-DKD data was not pre-specified and questionable statistically

ERG: Additional trial data from FIGARO-DKD and FIDELITY pooled analysis would provide insight into marketing authorisation population and reduce uncertainty

Committee: FIDELIO-DKD, FIGARO-DKD, and FIDELITY evidence are all relevant and appropriate for decision making to reduce uncertainty, even though marketing authorisation population represents approximately 90% of FIDELIO-DKD

ICER impact: decreases

model structure

Use of time-invariant transition probabilities in Markov model

Company: Time-invariant transition probabilities are appropriate; experts advised against more complex time-varying approach; method is common in CKD modelling and validated against SHARP-CKD-CVD model

ERG: Time-invariant approach is oversimplified; tight confidence intervals reflect this simplification; comparison with trial data over time would be more informative; ranges in SHARP-CKD-CVD represent extremes not confidence intervals

Committee: Effects of time-invariant transitions are uncertain; comparison of model predictions with Kaplan-Meier curves from trial data would be informative; external validation did not represent true external validation

ICER impact: uncertain_direction

model structure

How to model previous cardiovascular disease history in baseline population (45.9% of FIDELIO-DKD participants had prior CVD event)

Company: Use simplifying assumption that no patients had cardiovascular events before entering model; model from point of FIDELIO-DKD entry only

ERG: Model using total patient history; suggested ideal approach of 3 sub-models reflecting different CVD history states

Committee: Neither approach optimal; company approach likely resulted in optimistic cost-effectiveness; restructuring into 3 sub-models would reduce uncertainty. Company's updated sub-models did not properly track patients over time in each subgroup as intended

ICER impact: decreases

other

Sensitivity analysis methodology - parameter uncertainty and sampling approach

Company: Acknowledged limitations in original analyses; updated transition probabilities for standard care arm to be sampled from Dirichlet distribution

ERG: Original approach had multifaceted issues (grouping parameters, wide bounds, user-specified limits, utility overestimation); questioned why finerenone arm transitions not sampled from Dirichlet distribution like standard care

Committee: Updated approach was improvement but outputs remained uncertain; results should be interpreted with caution. Transition probabilities for CKD progression were critical and not subject to sensitivity analysis initially

ICER impact: uncertain_direction

other

Missing comparative effectiveness evidence with SGLT2 inhibitors

Company: No head-to-head comparison provided

ERG: Outstanding gap requiring acknowledgement

Committee: Noted as missing comparison; could not recommend finerenone instead of SGLT2 inhibitors, only as add-on; recommended alongside SGLT2 inhibitors

ICER impact: uncertain_direction

population generalisability

Relevance of trial populations defined by eGFR thresholds to marketing authorisation population

Company: Marketing authorisation population represents approximately 90% of FIDELIO-DKD population; trial enrolled only eGFR 25 ml/min/1.73 m2 and above

ERG: Concerned about lack of clarity regarding finerenone use when eGFR is 15-25 ml/min/1.73 m2 (CKD stage 4)

Committee: Marketing authorisation eGFR ranges are appropriate despite not covering all NHS CKD stage definitions; eGFR ranges in marketing authorisation are appropriate for likely finerenone use

ICER impact: uncertain_direction

stopping rule

Whether finerenone should be stopped when eGFR drops below 15 ml/min/1.73m² and after renal replacement therapy starts

Company: Finerenone stopped after renal replacement therapy starts

ERG: No strong preference

Committee: Finerenone stopped after renal replacement therapy starts

ICER impact: decreases

survival extrapolation

Whether health state-transition probabilities should remain time-invariant beyond the 4-year trial period when modelling over 34 years

Company: Time-invariant transition probabilities appropriate; validation against FIDELIO-DKD supported the approach

ERG: Modelling predictions for time to events should be explicitly compared with empirical Kaplan-Meier curves from trial; concerns about validity over 30 years of assumed fixed probabilities

Committee: Updated transition probabilities were also uncertain; acknowledged company addressed comparison with trial data but noted uncertainty remains for 30-year extrapolation beyond 4-year trial period

ICER impact: uncertain_direction

treatment effect duration

Whether treatment effects of finerenone combined with SGLT2 inhibitors are additive

Company: Effects of finerenone and SGLT2 inhibitors are independent and additive, based on evidence that background SGLT2 inhibitor use did not reduce finerenone benefit

ERG: Insufficient details provided; uncertain whether effects are truly additive versus additional but not independent

Committee: Uncertain; analyses provide a useful upper bound estimate but true additivity is not known

ICER impact: increases

treatment effect waning

Whether treatment benefit of finerenone persists beyond 4-year trial follow-up

Company: No treatment waning assumed because relative effect was almost constant over 4 years in trial; company assumes finerenone would be stopped if no treatment effect in practice

ERG: Not explicitly stated

Committee: Treatment effects beyond 4 years are uncertain; company provided scenario analyses with waning effects over 16 years but outcomes of extrapolation beyond 4 years are not known

ICER impact: decreases

treatment effect waning

Whether finerenone's treatment effect diminishes after the trial period and over what timeframe

Company: No treatment effect waning required; relative effect constant over 4-year trial; would be stopped in practice if no effect observed

ERG: Scenario analyses with 16-year waning considered but arbitrary; 7-year and 9-year discontinuation scenarios potentially more informative

Committee: Uncertainty around treatment waning inherent beyond trial period; company made reasonable attempt to explore; extrapolation beyond 4 years uncertain

ICER impact: decreases

utility source

Source of health state utility values for CKD stages

Company: Initially used FIDELIO-DKD trial-based utilities; revised after technical engagement to use TA358 (tolvaptan) utilities for consistency with accepted NICE approach

ERG: Preferred modified trial-based utilities over TA358 utilities which were from 2005 small population without EQ-5D; noted lack of engagement with more recent literature from NG28, NG148, NG203

Committee: Both approaches have advantages and disadvantages; during consultation company updated dialysis, transplant and cardiovascular event utilities to reflect NG28 values; ERG accepted updated utilities as appropriate

ICER impact: uncertain_direction

Evidence gaps

no direct comparisonNo direct comparison of finerenone against SGLT2 inhibitors when used as an add-on to standard care without SGLT2 inhibitors
short follow upLimited data on finerenone use when eGFR falls below 15 ml/min/1.73 m² (end-stage CKD)
short follow upFIDELIO-DKD trial follow-up limited to 4 years; cost-effectiveness model assumes lifetime horizon of 34.2 years with treatment effects beyond 4 years uncertain and based on time-invariant assumptions
no direct comparisonNo head-to-head comparison between finerenone plus SGLT2 inhibitors versus SGLT2 inhibitors alone; additivity of effects uncertain
single arm evidence onlyPrimary composite outcome in FIDELIO-DKD underpowered for individual components when restricted to marketing authorisation population; only one component statistically significant
no direct comparisonFinerenone has not been compared with SGLT2 inhibitors; can only be recommended as add-on rather than alternative to SGLT2 inhibitors
short follow upModel assumes time-invariant transition probabilities for 30 years beyond the 4-year trial data; treatment effects beyond 4 years uncertain
otherModelling of previous cardiovascular disease history uncertain; 45.9% of FIDELIO-DKD participants had prior CVD events but baseline assumption did not fully capture this

Special considerations

Innovation acknowledged Equality issues raised

Cross-references

TA775comparator guidance — NICE guidance on dapagliflozin for treating chronic kidney disease, referenced regarding SGLT2 inhibitor recommendations