FDA Endpoints and Noninferiority Margins for Antibacterial Approvals in cUTI and CABP

FDA endpoints and noninferiority margins for antibacterial approvals in cUTI and CABP

FDA's accepted efficacy standards for complicated urinary tract infection (cUTI) and community-acquired bacterial pneumonia (CABP) differ in structure. For cUTI, FDA recommends a composite of clinical and microbiological response in a microbiologically defined population, with a 10-percentage-point noninferiority (NI) margin ⁵⁹⁶⁰. For CABP, it recommends an early, symptom-based clinical response endpoint with a 12.5-percentage-point margin ⁷⁰⁷². Approvals since 2010 show where FDA has gone beyond those defaults. Several cUTI programs met a 15-point margin ²⁷²²⁰¹⁴²⁶⁵⁵. Recent CABP programs have generally used early clinical response at 72 to 120 hours in the ITT population, with margins as tight as 10 points ²⁸¹²⁸⁴⁵³⁹.

Key takeaways

cUTI default: composite clinical cure plus microbiological eradication in the micro-ITT population, assessed after the full treatment course plus at least 5 days of follow-up, with a 10% margin ⁵⁹⁶⁰.
cUTI flexibility: FDA's unmet-need guidance cites 15% as a potentially acceptable margin for patients with limited or no alternatives ⁸⁰. Vabomere, Zemdri, Fetroja and Zaynich were each approved on trials designed around a 15% margin ²⁷²²⁰¹⁴²⁶⁵⁵.
CABP default: improvement in CABP symptoms on day 4, with a 12.5% margin that rests on an M1 of at least 20 percentage points ⁷⁰⁷²⁷³.
CABP practice: Xenleta, Nuzyra and Baxdela all report early clinical response (ECR) at 72 to 120 hours in all randomized (ITT) patients ²⁸¹³⁰⁰⁶⁴⁹. Xenleta's Trial 2 and the Nuzyra CABP trial reviewed prespecified a 10% margin ²⁸⁴⁵³⁹.
Endpoint mismatch risk: when a legacy trial used an older test-of-cure (TOC) endpoint, FDA re-analyzed it against an early-response endpoint. Teflaro and Zevtera are the examples ³¹⁵⁵⁹⁷.

FDA guidance framework

cUTI: 2018 final guidance

FDA's final guidance Complicated Urinary Tract Infections: Developing Drugs for Treatment is dated June 1, 2018 ⁴²⁶⁰⁵. Its core recommendations:

Primary endpoint: a responder endpoint that combines microbiological success with clinical resolution in patients with microbiologically documented cUTI ⁵⁹.
Primary population: micro-ITT, with consistency checked across the other analysis populations ⁵⁹⁶⁰.
Timing: a fixed time after randomization that covers the full course of antibacterial therapy plus at least 5 days of post-treatment observation. The NI margin is 10 percentage points ⁵⁹.
IV-to-oral switch designs: FDA also supports the composite endpoint at about 5 days of IV therapy (the switch point), again with a 10-point margin. Patients should still be followed through the end of treatment and at least 5 days beyond, for example to about day 14 after randomization ⁶¹.
Margin derivation (post-treatment endpoint): FDA compared an upper bound of about 33% success without antibacterial therapy against a lower bound of about 70% from modern trials. That gives an estimated 37-point treatment effect, which FDA discounted to M1 = 30% before setting M2 = 10% ⁶²⁶³.
Margin derivation (end-of-IV endpoint): the comparison was 33% versus 56%, a 23-point effect discounted to M1 = 20%, with M2 = 10% chosen on clinical grounds ⁶²⁶³.
Larger margins: sponsors should discuss with FDA any proposed margin above 10% ⁶²⁶³.

CABP: 2014 final guidance

FDA's final guidance Community-Acquired Bacterial Pneumonia: Developing Drugs for Treatment is dated January 10, 2014 ⁶⁴⁶⁰⁴.

Primary endpoint: clinical response based on improvement in CABP symptoms (chest pain, cough frequency or severity, productive sputum, and difficulty breathing), assessed on day 4 after enrollment ⁷⁰.
Margin: FDA considers 12.5 percentage points generally reasonable for the day-4 clinical-response endpoint. A larger margin may be considered in some circumstances, but sponsors should discuss it with FDA before the trial starts ⁷².
Margin justification: FDA judged an M1 of at least 20 percentage points to be a conservative estimate of the active-comparator effect on symptom improvement around days 3 to 5. That estimate supports an M2 of 12.5 points, and FDA notes that choosing M2 calls for clinical judgment and sponsor justification ⁷³⁶⁵.
Analysis population: the primary population should be agreed with FDA in advance. ITT (all randomized patients) is generally acceptable ⁶⁹⁹⁷⁰⁰. For a narrow-spectrum drug, such as one active against a single genus and species, micro-ITT is the primary population ⁷¹.
Secondary endpoints and follow-up: secondary endpoints include day-4 response, clinical outcome at end of therapy, and symptom resolution at a fixed time after therapy ⁷⁰¹. The guidance calls for a post-therapy visit 5 to 10 days after treatment and an all-cause mortality assessment at day 28 ⁷⁰². It also describes 28-day all-cause mortality as a possible primary endpoint, subject to feasibility and agreement with FDA ⁷⁰³.

The 2024 ceftobiprole (Zevtera) review cites a 2020 FDA CABP guidance whose primary endpoint is day-4 improvement in at least two of chest pain, cough, productive sputum and difficulty breathing, with no worsening ⁶⁰¹.

Unmet-need and limited-population flexibility

Larger margins for limited-option populations: under Antibacterial Therapies for Patients With an Unmet Medical Need for the Treatment of Serious Bacterial Diseases (2017), a margin larger than the disease-specific default may be acceptable for a drug intended for patients with limited treatment options. Acceptance depends on the type and degree of unmet need ⁷⁸⁷⁹⁶⁰⁷.
cUTI example: the guidance cites a margin above 10% but below the roughly 30% estimated treatment effect, such as 15%, as potentially acceptable ⁸⁰.
No CABP figure: the retrieved guidance text gives no comparable CABP-specific figure ⁷⁷.
2025 Q&A: the June 2025 Q&A guidance repeats that a randomized NI trial against best available therapy, supported by confirmatory evidence, can establish efficacy where unmet need remains ⁶⁸⁸⁶¹³.
LPAD pathway: FDA describes NI trials with wider-than-traditional margins as one possible streamlined approach ⁸¹.

cUTI approvals: endpoints, margins and results

Product	Pivotal trial and comparator	Primary endpoint, timing, population	NI margin	Result (difference, 95% CI)
Zerbaxa (ceftolozane/tazobactam)	Pooled CXA-cUTI-10-04/-05 vs levofloxacin ²⁵⁹²³²⁶⁰	Composite clinical cure and microbiological eradication at TOC, 7 ± 2 days after last dose, mMITT ²⁶¹²⁶²	10% ²²²⁶³	76.9% vs 68.4%; +8.5 (2.3 to 14.6) ²⁶⁴
Avycaz (ceftazidime/avibactam)	cUTI Trial 1 vs doripenem ²³⁷	Two primary endpoints in mMITT: symptomatic response at Day 5, and combined symptomatic response plus microbiological cure at TOC (21 to 25 days after randomization) ²³⁷	Not stated in retrieved text ²²³	Day 5: +4.0 (−2.4 to 10.4). TOC: +6.7 (0.3 to 13.1) ²³⁸²³⁹
Vabomere (meropenem/vaborbactam)	Study 505 vs piperacillin/tazobactam ²⁰³²⁰⁴	Overall response (clinical cure or improvement plus eradication) at end of IV therapy, m-MITT ²⁷²⁰⁶	15% ²⁷²⁰⁶	98.4% vs 94.0%; +4.5 (0.7 to 9.1) ²⁹²⁰⁷
Zemdri (plazomicin)	Study 009 vs meropenem ²¹⁸	Co-primary composite cure at Day 5 and at TOC (Day 17 ± 2), mMITT ²²⁰	15% at both visits ²²⁰	Day 5: −3.4 (−10.0 to 3.1). TOC: +11.6 (2.7 to 20.3) ²²¹
Fetroja (cefiderocol)	Study 1409R2121 vs imipenem/cilastatin, 2:1 ¹³⁸¹³⁹	Composite clinical response plus eradication at TOC, 7 ± 2 days after end of treatment, Micro-ITT ¹⁴²¹⁴³¹⁴⁴	15% ¹⁴²¹⁴³¹⁴⁴	72.6% vs 54.6%; +18.6 (8.2 to 28.9) ¹⁴⁵
Recarbrio (imipenem/cilastatin/relebactam)	Phase 2 PN003 vs imipenem/cilastatin plus placebo ⁵⁷⁶⁵⁶⁰	FDA assessed combined clinical and microbiological response (<10³ CFU/mL) at early follow-up in MITT ⁵⁷⁸⁵⁷⁹⁵⁸⁰	15% design margin. FDA noted 10% as the guidance standard ⁵⁸¹	Not shown noninferior. Indication approved on limited clinical data ⁵⁸²⁵⁸³
Exblifep (cefepime/enmetazobactam)	AT-301 vs piperacillin/tazobactam ⁵⁴⁹⁵⁵⁰	Overall success (clinical cure plus eradication, <10³ CFU/mL) at TOC, m-MITT ⁵⁵¹	10% ⁵⁵²	79.1% vs 58.9%; +20.2, lower 95% bound 14.3 ⁵⁵²
Contepo (fosfomycin IV)	Trial 1 vs piperacillin/tazobactam ⁶³⁵	Overall success at TOC, 19 (+2) days after randomization, mMITT ⁶³⁶	Not stated in retrieved text ⁶³⁷	63.5% vs 55.6%; +7.9 (−3.1 to 18.9) ⁶³⁷
Utebzi (tebipenem pivoxil, oral)	SPR994-305 vs IV imipenem/cilastatin ⁶⁶⁰	Overall response at TOC (Day 17 ± 2), micro-ITT ⁶⁶⁰	10% ⁶⁶⁰	58.5% vs 60.2%; −1.3 (−7.5 to 4.8) ⁶⁶⁰
Zaynich (cefepime/zidebactam)	Phase 3 vs meropenem ⁶⁵⁵	Overall success at TOC, mMITT-STIC ⁶⁵⁵	15% ⁶⁵⁵	89.0% vs 68.4%; +20.6, lower 95% bound 12.3 ⁶⁵⁵

Patterns in the cUTI record

The 10% and 15% margins coexist. The guidance default of 10% applies to Zerbaxa, Exblifep and Utebzi ²²⁵⁵²⁶⁶⁰. A 15% margin was prespecified and met for Vabomere, Zemdri, Fetroja and Zaynich ²⁷²²⁰¹⁴²⁶⁵⁵. This matches the unmet-need guidance's example of 15% for cUTI ⁸⁰.

Timing has moved toward fixed days after randomization. Zerbaxa and Fetroja assessed TOC 7 ± 2 days after the last dose ²⁶¹¹⁴². Other programs fixed the visit relative to randomization or first dose:

Day 17 ± 2 for Zemdri and Utebzi ²²⁰⁶⁶⁰
19 (+2) days for Contepo ⁶³⁶
21 to 25 days for Avycaz ²³⁷

This is consistent with the guidance's recommendation of a fixed time point after randomization ⁵⁹.

Early-timepoint endpoints for IV-only products. Vabomere's primary endpoint was at end of IV therapy ²⁷. Avycaz and Zemdri paired a Day 5 endpoint with a later TOC endpoint ²³⁷²²⁰. The Vabomere review notes that FDA guidance recommended both end of IV therapy and TOC as co-primary endpoints for IV-only drugs ²⁰⁹. FDA's own Vabomere analysis also reported TOC results: 76.5% vs 73.2%, difference 3.3 (95% CI −6.2 to 13.0) ²⁰⁸.

Analysis-population choices can decide the outcome.

For Vabomere, FDA excluded two study sites over data-quality concerns. Its analysis population therefore differed from the applicant's, though noninferiority still held (difference 4.1, 95% CI 0.3 to 8.8) ²⁰⁸.
For tebipenem, FDA found that the earlier trial SPR994-301 did not meet its NI criterion when restricted to patients with ertapenem-susceptible pathogens. That finding led to a complete response letter and to the confirmatory SPR994-305 trial ⁶⁶¹⁶⁶².
In the Avycaz phase 2 review, FDA treated mMITT, not the applicant's ME population, as primary ²⁴⁴.

Microbiological thresholds vary. Most programs defined eradication as reducing uropathogens to <10⁴ CFU/mL ²⁶¹²⁰⁶¹⁴²⁶³⁶. Exblifep used <10³ CFU/mL ⁵⁵¹, and FDA's Recarbrio assessment also used <10³ CFU/mL ⁵⁷⁸.

A composite driven by microbiology. In the Zaynich review, FDA observed that microbiological success largely drove the composite result. It considered noninferiority established but the interpretation of superiority uncertain ⁶⁵⁶.

CABP approvals: endpoints, margins and results

Product	Pivotal trial(s) and comparator	Primary endpoint, timing, population	NI margin	Result (difference, 95% CI)
Teflaro (ceftaroline fosamil)	P903-08 and P903-09 vs ceftriaxone ³¹²	Applicant: clinical cure at TOC in co-primary CE and MITTE populations ³¹³³¹⁴. FDA: Day 4 (72 to 96 h) response in microbiologically confirmed patients ³¹⁵³¹⁶	Applicant 10% at TOC ³¹³³¹⁴. FDA 10% for the Day 4 analysis ¹⁰³	FDA Day 4: +11.2 (−4.6 to 26.5) and +7.6 (−6.8 to 21.8) ¹⁰⁴¹⁰⁵
Xenleta (lefamulin)	Trial 1 (3101) vs moxifloxacin with or without linezolid. Trial 2 (3102) vs oral moxifloxacin ²⁷⁸²⁷⁹²⁸⁰	ECR 72 to 120 h after first dose, ITT ²⁸¹²⁸²	10% for Trial 2, reduced from 12.5% by protocol amendment ²⁸⁴	Trial 1: −2.9 (−8.5 to 2.8). Trial 2: 0.0 (−4.4 to 4.5) ²⁸⁶²⁸⁷
Nuzyra (omadacycline)	Trial 1 and Trial 4 vs moxifloxacin ²⁹⁹	ECR 72 to 120 h, ITT ³⁰⁰³⁰¹	10% prespecified for the CABP trial reviewed ⁵³⁹⁵⁴⁰	Trial 1: −1.6 (−7.1 to 3.8). Trial 4: +1.9 (−3.0 to 6.8) ³⁰⁴³⁰⁵
Baxdela (delafloxacin)	Trial 3 vs moxifloxacin (switch to linezolid if MRSA) ⁶⁴⁸	ECR 72 to 120 h, ITT ⁶⁴⁹	Not stated in retrieved text ⁶⁴⁰	−0.2 (−4.4 to 4.1) ⁶⁵⁰
Zevtera (ceftobiprole medocaril)	CAP-3001 vs ceftriaxone with or without linezolid ⁵⁹³	Original: clinical cure at TOC ⁵⁹⁶. Re-analysis for the US submission: clinical success at Day 3, ITT ⁵⁹⁷⁵⁹⁸	Original 15% at TOC ⁵⁹⁶. FDA sensitivity analysis showed NI within 10% ⁵⁹⁹⁶⁰⁰	Day 3 ITT: −0.1 (−7.2 to 7.1) ⁵⁹⁷⁵⁹⁸

Patterns in the CABP record

ECR at 72 to 120 hours in ITT is the operative primary endpoint. Xenleta, Nuzyra and Baxdela define responders in similar terms ²⁸¹³⁰⁰⁶⁴⁹:

the patient survived
at least two of cough, sputum production, pleuritic chest pain and dyspnea improved
none of those symptoms worsened
no rescue antibacterial was given (Nuzyra also excludes discontinuation for an adverse event ³⁰¹)

In the Xenleta review, FDA kept the ITT ECR endpoint and added a worst-case analysis of indeterminate responses. Noninferiority held under both the applicant's and the worst-case analyses ²⁸²²⁸³²⁸⁸²⁸⁹.

Margins tighter than the guidance default. The guidance supports 12.5% ⁷². Xenleta's Trial 2 tightened its margin from 12.5% to 10% by protocol amendment ²⁸⁴. The Nuzyra review records a prespecified 10% margin while noting the 12.5% guidance figure ⁵³⁹⁵⁴⁰.

Dual endpoints for FDA and EMA. Xenleta's EMA-directed endpoint was investigator-assessed clinical response at TOC in a modified ITT population, with a 10% margin ²⁸⁴²⁸⁵. Nuzyra reported investigator-assessed response at post-treatment evaluation as 87.6% vs 85.1%, difference 2.5 (95% CI −2.4 to 7.4) ⁵⁴¹. Baxdela reported investigator-assessed TOC success of 90.5% vs 89.7% (difference 0.8, 95% CI −3.3 to 4.8) ⁶⁵¹⁶⁵².

Legacy TOC trials were re-analyzed.

Teflaro: FDA did not accept the applicant's 10% margin for clinical cure at TOC because it lacked adequate historical justification. FDA instead analyzed Day 4 response in microbiologically confirmed patients against a 10-point margin, and both trials met it ³¹⁵³¹⁶¹⁰³.
Zevtera: CAP-3001 was originally run with a TOC endpoint and a 15% margin, then re-analyzed for the US submission using Day 3 clinical success ⁵⁹⁶⁵⁹⁷. The review notes that this Day 3 measure differs from the guidance's Day 4 symptom definition in timing and symptom scoring ⁶⁰¹⁶⁰². The reviewer judged CAP-3001 adequate provided it was supported by positive findings in the program's SAB and ABSSSI indications ⁶⁰³.

Practical implications for development programs

cUTI margin: plan around 10% unless the target population has a documented unmet need. Programs that met 15% enrolled general cUTI populations, though Zemdri's and Fetroja's labels were limited to patients with limited or no alternative treatment options. FDA's guidance asks sponsors to discuss any margin above 10% in advance ⁶²⁶³⁸⁰.
cUTI timing: fix the TOC visit relative to randomization, and expect FDA to look closely at the micro-ITT population, including subsets defined by comparator susceptibility ⁵⁹⁶⁶¹.
IV-only products: if a product has no oral step-down, an early (Day 5 or end of IV) endpoint paired with TOC fits both the guidance and recent precedent ⁶¹²²⁰²⁰⁹.
CABP endpoint and margin: design around ECR at 72 to 120 hours in ITT. Recent precedent supports 10%, which is tighter than the 12.5% guidance figure, and an investigator-assessed TOC endpoint is often added for ex-US regulators ²⁸¹²⁸⁴⁵³⁹²⁸⁵.
Legacy CABP trials: if a trial used a TOC-only design, budget for an FDA re-analysis on an early-response endpoint. Retrospective endpoint definitions may not match the guidance exactly ³¹⁵⁶⁰¹⁶⁰².

Open questions to explore further

The retrieved FDA records did not state the cUTI NI margins for Avycaz or Contepo, or the CABP margin for Baxdela ²²³⁶³⁷⁶⁴⁰. Reviewing those approvals' statistical reviews directly would close the gaps. Other useful follow-ups:

how FDA handled HABP/VABP all-cause mortality margins in the same programs
how pediatric extrapolation from adult cUTI and CABP data has been accepted
whether the 2020 CABP guidance version cited in the Zevtera review changed the 12.5% margin ⁶⁰¹