type: framework-update tags: [specialty-pharma, fda, pdufa, surrogate-endpoint, primary-endpoint, rare-disease, approval-probability, unmet-medical-need] confidence: medium created: 2026-04-02 source: TVTX stock-analysis 2026-04 persona: wsm provenance: legacy source_analysis_path: null source_paragraph_quote: null source_transcript_span: null source_loss_log_path: null

Imperfect-Trial PDUFA: Approval Probability When Primary Endpoint Is Missed

When a Phase 3 trial misses its primary hard endpoint (e.g., eGFR slope in kidney disease) but hits a validated surrogate (e.g., proteinuria reduction), the PDUFA approval probability cannot be read from the trial result alone. It requires combining four signals: (1) AdCom removal — FDA resolved questions internally; (2) unmet medical need — first-in-class in a disease with zero approved therapies lowers the bar; (3) regulatory working-group surrogate validation — a pre-established framework endorsing the surrogate as a meaningful endpoint; (4) post-hoc data on hard outcomes — even if not the primary endpoint, directional benefit on ESRD/mortality shifts the risk-benefit calculus. Absent all four mitigants, a missed primary is near-terminal for approval. With all four, probability reaches 60-65% — higher than a coin flip but not a slam dunk.

Evidence

• TVTX FILSPARI FSGS PDUFA (April 13, 2026): DUPLEX trial missed primary eGFR slope endpoint (0.3 ml/min/1.73m²/year difference, 95% CI -1.7 to 2.4; not significant). Hit surrogate: 64.7% partial proteinuria remission vs 43.9% irbesartan (statistically significant).
• Mitigants present: (a) AdCom removed after Major Amendment submission; (b) FSGS has zero approved therapies; (c) PARASOL nephrology workgroup validated proteinuria as meaningful FSGS endpoint; (d) post-hoc data showed lower ESRD rates and higher complete remission at 108 weeks.
• WSM probability estimate: 60-65% approval vs ~50% baseline for a missed-primary sNDA.
• The "active comparator design" (vs irbesartan, a standard-of-care ARB) made the primary endpoint harder to achieve — smaller delta vs placebo designs — which is a mitigant in the interpretation of the miss.

Implication

Build a checklist for imperfect-trial PDUFA assessment. Start at 45-50% base rate for any missed-primary submission. Add probability mass for each present: AdCom removed (+10-15pp), first-in-class unmet need (+5-10pp), regulatory-endorsed surrogate (+5pp), directional post-hoc hard-outcome data (+5pp), active comparator design explaining the miss (+5pp). Cap at ~70% — missed-primary is always a genuine FDA concern regardless of mitigants. If multiple mitigants absent (e.g., AdCom convened AND no unmet need), revise down to 30-40%. Always flag a Major Amendment extension as a yellow flag (FDA wanted more data) rather than a red flag — by itself it does not change the base rate materially.