FDA-Accepted Primary Endpoints and Pivotal Trial Designs in SLE and Lupus Nephritis Approvals

FDA-accepted primary endpoints and pivotal trial designs in SLE and lupus nephritis approvals

FDA's modern targeted approvals for systemic lupus erythematosus (SLE), since 2011, have rested on two composite responder endpoints: the SLE Responder Index (SRI-4) for belimumab and the BILAG-based Composite Lupus Assessment (BICLA) for anifrolumab. Both were measured at Week 52 in placebo-controlled trials that added the study drug to standard therapy. For lupus nephritis (LN), every adult approval has rested on a renal response composite: proteinuria (UPCR) plus an eGFR preservation criterion, with steroid-use and rescue-therapy conditions attached. The assessment timepoint has varied from 52 to 104 weeks. The sections below set out what each approval package used, how FDA reviewers viewed the endpoints, and what that means for development programs.

Approvals at a glance

Systemic lupus erythematosus

Common design architecture

All SLE approvals share one basic design. Patients were randomized to the study drug or placebo, each added to standard therapy. Efficacy was measured at Week 52, and the population had active, autoantibody-positive disease.

• Add-on to standard therapy. In the belimumab IV trials, belimumab 1 or 10 mg/kg or placebo was added to standard therapy. Common baseline treatments were corticosteroids, immunosuppressives and antimalarials ⁵²²⁵⁰². The anifrolumab phase 3 trials used the same background classes and stratified randomization by baseline disease severity, Day 1 oral corticosteroid (OCS) dose and interferon gene-signature status ³⁴⁸³⁵².
• Enrichment for active, seropositive disease. BEL112341 (belimumab SC) required ACR-criteria SLE, a screening SELENA-SLEDAI of at least 8, and a positive ANA and/or anti-dsDNA test. It excluded severe active LN and severe active CNS lupus ³²¹.
• Randomization and size. BEL112341 was multicenter, randomized, double-blind and placebo-controlled. It randomized 836 participants 2:1 to belimumab 200 mg SC weekly or placebo ³¹⁹³²⁰. TULIP-1 randomized 457 patients 1:2:2 to anifrolumab 150 mg, 300 mg or placebo ³⁴⁸. TULIP-2 randomized 362 patients 1:1 ³⁵⁰.
• Protocol-mandated steroid tapering. In both anifrolumab phase 3 trials, patients on at least 10 mg/day OCS at baseline had to taper to no more than 7.5 mg/day during Weeks 8 to 40 unless disease activity worsened ³⁴⁸³⁵⁴. The SC anifrolumab trial used the same taper rule ⁵³⁶.
• Strict non-responder imputation. For BICLA, patients who discontinued treatment or used restricted medications beyond protocol thresholds counted as non-responders ³⁵³. The SC anifrolumab trial added death to these rules ⁵³⁸.

SRI-4: the belimumab endpoint

SRI-4 response at Week 52 requires all three of the following ⁵²³³²²:

1. At least a 4-point reduction in SELENA-SLEDAI from baseline.
2. No new BILAG A organ-domain score and no more than one new BILAG B score.
3. No worsening in Physician's Global Assessment (an increase of less than 0.30 points).

In the adult IV program, 10 mg/kg was statistically superior to placebo in both phase 3 trials. The 1 mg/kg dose was not consistently superior ⁵²⁴, and it is not recommended ⁶³⁹. The SC program relied on a single phase 3 trial, which met its primary endpoint ³¹⁸⁵²⁷. Its key secondary endpoints were time to first severe flare and steroid reduction in patients taking more than 7.5 mg/day prednisone at baseline. Steroid reduction was defined as an average dose cut of at least 25% to no more than 7.5 mg/day over Weeks 40 to 52. The severe-flare endpoint was significant, but the steroid-reduction endpoint was not (p=0.0732) ³²³.

BICLA: the anifrolumab endpoint and the TULIP endpoint switch

BICLA response at Week 52 requires improvement in every organ domain that had moderate or severe activity at baseline: BILAG A scores must fall to B, C or D, and BILAG B scores to C or D. It also requires all of the following ³⁵³:

• no new BILAG A score and no two new BILAG B scores
• no increase in SLEDAI-2K
• no PGA worsening of 0.30 points or more
• no treatment discontinuation
• no restricted medication use beyond protocol thresholds

The anifrolumab program is the main precedent on how FDA handles a change of primary endpoint in SLE:

• TULIP-1 missed its SRI-4 primary endpoint. Responses with 300 mg were 36.1% vs 40.2% with placebo (p=0.41) ³⁴⁹. Reviewers treated its favorable BICLA result as exploratory or nominal ³⁸⁹. Revising the medication rules after the fact did not make its SRI-4 result nominally significant ³⁹⁰.
• TULIP-2 changed endpoints before unblinding. After TULIP-1, the sponsor changed TULIP-2's primary endpoint from SRI-4 to BICLA, with the Division's agreement, before database lock and unblinding ³⁹¹³⁹². TULIP-2 met the BICLA endpoint, and its Week 52 SRI-4 result also favored anifrolumab ³⁹³.
• FDA considered both endpoints valid. Reviewers judged SRI-4 and BICLA both to be valid, clinically meaningful SLE response measures that can classify the same patient differently. SRI-4 requires a 4-point SLEDAI-2K improvement and can miss partial improvement. BICLA can capture substantial partial improvement across affected organs ³⁸⁶³⁸⁷.
• Evidentiary basis. TULIP-2 was the adequate and well-controlled trial providing primary evidence of effectiveness. TULIP-1 and the phase 2 study did not count independently as successful trials, but together they supplied confirmatory evidence. The division director also cited a supportive mechanistic rationale ³⁹⁴³⁹⁵.
• Integrity checks. FDA asked for analyses under the original rules, the amended rules, and observed outcomes regardless of rule adherence or discontinuation. Reviewers found consistent TULIP-2 BICLA and SRI-4 effects under both rule sets. After reviewing monitoring-board minutes, they saw no sign that the endpoint change compromised trial integrity ^379393396.

The phase 2 study 1013 used SRI-4 with OCS tapering at Week 24 as its primary endpoint. Reviewers considered its significance nominal because the dose-ranging study lacked appropriate multiplicity control ³⁸⁸.

The SC anifrolumab approval kept BICLA at Week 52 as the primary endpoint. Its primary analysis was an interim analysis with a prespecified alpha of 0.0354 ^537539540.

Pediatric SLE: extrapolation from adult data

The pediatric belimumab indication rested on PLUTO, a randomized, double-blind, placebo-controlled 52-week trial in 93 children aged 5 to 17. It used the same SRI-4 Week 52 endpoint as the adult trials ^455456458. PLUTO was too small for formal hypothesis testing on its own ⁴⁵⁹.

FDA based its conclusion on several lines of evidence ^460461462:

• the established adult efficacy from studies C1056 and C1057
• comparable dosing and pediatric drug exposure
• a pediatric response pattern consistent with the adult results
• a post-hoc Bayesian analysis that borrowed from the adult data

Lupus nephritis

Renal response definitions compared

Belimumab also tested a stricter Complete Renal Response endpoint (uPCR <0.5 g/g, plus either eGFR ≥90 or no >10% decrease in eGFR) as a ranked secondary endpoint. It reached 30% vs 20% at Week 104 (P=0.017) ⁴²¹⁴²⁴. PERR at Week 52 was also significant: 47% vs 35% (P=0.025) ⁴²⁵. Note that UPCR thresholds in mg/mg and g/g are the same ratio, so 0.5 mg/mg and 0.5 g/g are the same cutoff.

Design features

• Background therapy is part of the design. Lupkynis was tested on top of mycophenolate mofetil (MMF, mean 1.9 g/day in AURORA-1) and corticosteroids. Steroids began with IV methylprednisolone on Days 1 and 2, followed by oral prednisone tapered toward 2.5 mg/day by Week 16 ⁴⁹⁶⁴⁹⁷. REGENCY added obinutuzumab or placebo to MMF 2 to 2.5 g/day and tapering corticosteroids starting at 0.5 mg/kg/day ³⁶⁴³⁶⁵. BLISS-LN allowed MMF or cyclophosphamide induction ⁴¹⁹.
• Biopsy-confirmed, proteinuric populations. AURORA-1 enrolled patients with biopsy-proven Class III/IV LN and UPCR ≥1.5 mg/mg, or Class V with UPCR ≥2.0 mg/mg. It excluded eGFR ≤45 ⁴⁹⁰⁴⁹¹. REGENCY enrolled Class III or IV LN (with or without Class V) with UPCR ≥1 g/g and excluded eGFR <30 ³⁶⁶.
• Number of trials. Lupkynis was approved on one phase 3 trial (AURORA-1), supported by a phase 2 trial. AURA-LV's primary endpoint was CRR at Week 24, and the 23.7 mg BID dose was superior to placebo there (32.6% vs 19.3%, p=0.045) ⁴⁸⁸⁴⁹³. Benlysta's LN indication rested on BLISS-LN, a 104-week double-blind, placebo-controlled trial in 448 adults ⁶⁵⁸.
• Key secondary endpoints. AURORA-1 used time to UPCR ≤0.5 mg/mg, partial renal response (≥50% UPCR reduction) and time to 50% UPCR reduction ⁴⁹⁸. REGENCY used three: CRR with a successful prednisone taper (no more than 7.5 mg/day from Weeks 64 to 76), proteinuric response (UPCR <0.8 g/g), and renal-related events or death ³⁶⁷.

How FDA has viewed proteinuria-driven renal response

The Lupkynis review is the clearest statement of FDA's position on renal response endpoints:

• CRR was accepted as clinically meaningful. FDA treated near-normal proteinuria with preserved renal function, achieved without rescue treatment or excess steroid use, as a reasonable predictor of long-term benefit. It acknowledged that the link to end-stage renal disease and death has not been definitively validated ^470563564.
• Benefit-risk caveats. The deputy director found convincing statistical evidence that voclosporin increased CRR, mainly through reduced proteinuria. The deputy director also noted that proteinuria is a surrogate and that the size of the long-term clinical benefit relative to risk was unclear ⁴⁶⁴. Reviewers flagged a higher risk of decreased GFR ⁴⁶⁵.
• Labeling limitations reflect trial scope. Safety and efficacy beyond one year were not established ⁵⁶⁵. Use with cyclophosphamide is not recommended ⁵⁶⁷. Use is not recommended when baseline eGFR is ≤45 unless benefit outweighs risk, because those patients were excluded from the trials ⁵⁶⁸.

Subgroup results in BLISS-LN show the risk of a heterogeneous population. In the cyclophosphamide-induction subgroup, the PERR odds ratio was 1.5 with a confidence interval that included 1 ⁶⁵⁹⁶⁶⁰. In the Class V subgroup, PERR numerically favored placebo (36% vs 42%) ⁶⁶¹.

FDA guidance context

FDA's only lupus-specific guidance in the reviewed corpus is the 2010 final guidance Systemic Lupus Erythematosus: Developing Medical Products for Treatment ¹. It states that no SLE biomarker has been validated as a surrogate endpoint, so biomarkers cannot replace direct assessment of clinical benefit ⁵⁷⁴. When trials use reduction in disease activity as an endpoint, the guidance recommends trained, consistent disease-index scoring and possibly an adjudication committee ⁵⁷³.

The guidance also covers two specific designs:

• Steroid-sparing claims. It recommends an add-on design that starts during a flare with induction-dose steroids in both arms and then tapers. The efficacy measure is the proportion reaching ≤10 mg/day prednisone-equivalent with quiescent disease and no flares for at least three consecutive months in a one-year trial. A steroid-reduction claim should be paired with another clinical benefit, such as reduced disease activity, as the primary endpoint ^429430431.
• Serious acute manifestations. The primary endpoint should be improvement in the affected-organ score on a disease activity index. Durability and safety should be followed for at least one year in total ⁴³¹.

Practical implications for development programs

• SLE primary endpoint. Both SRI-4 and BICLA at Week 52 have supported approval, and FDA reviewers regard both as valid measures that capture different patterns of response ³⁸⁶³⁸⁷. Prespecify the alternative endpoint as a key secondary endpoint. Plan for discordance between the two, as TULIP-1 showed ³⁸⁷.
• Endpoint changes. A mid-program endpoint change can be acceptable if it is agreed with the Division and locked before unblinding. Expect FDA to ask for sensitivity analyses across medication-rule versions and for data monitoring committee records ³⁹¹³⁹⁶.
• Steroid handling. Mandated OCS tapers and restricted-medication non-responder rules are standard in recent SLE trials ³⁴⁸³⁵³. A steroid-reduction secondary endpoint can fail even when the primary endpoint succeeds, as it did in BEL112341 ³²³.
• LN primary endpoint. A renal response composite combining proteinuria, eGFR, steroid limits and rescue rules is the accepted standard. The timepoint has ranged from 52 weeks (Lupkynis) to 76 (Gazyva) and 104 weeks (Benlysta) ^492367420. Shorter trials can bring one-year limits into labeling ⁵⁶⁵.
• Single pivotal trial. Lupkynis (phase 3 plus phase 2) and Saphnelo (one successful phase 3 plus confirmatory evidence) show that one adequate and well-controlled trial with supportive data can be enough ⁴⁸⁸³⁹⁴.
• Pediatrics. Matched dosing, comparable exposure and the same primary endpoint allowed FDA to extrapolate adult SLE efficacy to children from a trial too small for hypothesis testing on its own ^459460461.