The case for collecting airway pharmacodynamic data in early-phase respiratory trials — and the consequence of not.
Every program lead in respiratory drug development has been in the room for the same conversation. The Phase II data has come in. The primary endpoint is equivocal. The biomarker package was systemic — blood eosinophils, serum cytokines, maybe FeNO if the program had Th2 biology to begin with. Someone asks the question that decides the next two years and several hundred million dollars of spend: did the drug fail to engage the target, or did the trial fail to measure target engagement where it mattered?
The honest answer is that nobody knows. The biology where the drug was supposed to act — the airway compartment — was never directly examined. The decision gets made anyway, because it has to be. Most of the time the program stops; sometimes it advances on indirect proxies. Either way, the analysis that might have distinguished a failed drug from a failed measurement is no longer available.
The decision to collect airway data, or not, is made once. By default if not by choice. And it is irreversible.
The real cost of not measuring where the drug acts
When you sit with a senior clinical development colleague long enough, the same pattern shows up. A weak Phase II read in a respiratory program produces a binary that isn’t actually binary. “The drug didn’t work” is one possibility. “The drug worked but we weren’t looking at the right biology” is another. Without lung-compartment pharmacodynamic data, those two failure modes are indistinguishable on paper, even though they have completely different downstream implications: one warrants stopping the program, the other warrants redesigning it.
The same gap shows up on the upside, less visibly. A clean positive Phase II read with no airway PD data still leaves the development team designing Phase III endpoints and stratification rules from systemic measures alone. That works — it has worked — but it leaves precision on the table. Responder analysis is harder. Dose-response modeling at the site of action is unavailable. Salvage analyses, if the Phase III primary endpoint slips, have nowhere to go.
None of this is news to anyone who has run a respiratory program. What is news, or at least worth naming, is that the tools to close the gap are operationally cheaper and methodologically more mature than most program teams realize.
Why exhaled breath condensate is the noninvasive option that fits
Exhaled breath condensate (EBC) is collected by having a patient breathe tidally through a chilled condenser for about ten minutes. Aerosolized airway lining fluid droplets condense on the cooled surface and are recovered. The resulting sample contains measurable concentrations of inflammatory mediators, oxidative stress markers, cytokines, leukotrienes, nucleic acids, pH, and volatile organic compounds — a noninvasive read on the airway compartment a bronchoscopy would access, repeatable at every visit, with no procedural risk.
BAL gives you the highest analytical fidelity but is impractical for serial use. Blood markers don’t reflect lung-compartment biology. Induced sputum is feasible but burdensome and the induction itself alters what you’re measuring. FeNO is excellent for Th2-driven eosinophilic inflammation but answers only that one question. EBC fits the gap: high lung specificity, no invasiveness, excellent serial feasibility, and access to the soluble mediator chemistry the others miss.
EBC biomarkers are exploratory under the FDA-NIH BEST framework, not formally qualified Drug Development Tools. That distinction matters. But “exploratory” in this context carries no requirement for statistical powering, no impact on primary endpoint strategy, and no added regulatory exposure — provided the data are pre-specified and clearly designated as exploratory in the protocol and SAP. Scope conditions, not disqualifying ones.
Example: What the data actually look like: Kazani et al., 2013
An example for what well-collected EBC can do comes from a study most respiratory development scientists know but may not have looked at recently. Kazani and colleagues, working within the NHLBI’s Severe Asthma Research Program (SARP) at Brigham and Women’s Hospital, used the RTube™ to collect EBC from 81 subjects across the asthma severity spectrum and a healthy control arm. They measured two opposing arms of arachidonic acid metabolism: pro-inflammatory leukotriene B4 (LTB4) and pro-resolving lipoxin A4 (LXA4).
EBC LTB4 distinguished asthma from health with 100% sensitivity and specificity — AUC 1.0, no overlap. EBC LXA4 reached AUC 0.96. The LXA4/LTB4 ratio fell 41% in severe versus moderate asthma (P=0.034) and correlated with FEV1 — a pathway-balance insight that severity reflects a relative deficit of pro-resolving mediators rather than simply an excess of pro-inflammatory ones, something single-analyte and cell-counting endpoints cannot resolve. In a head-to-head comparison within the same cohort, FeNO at the standard 25 ppb threshold misclassified 32% of moderate and 61% of severe asthmatics. The EBC cutoffs misclassified essentially none.
This is one cross-sectional study in one indication. It does not prove that EBC has changed a development decision. It does prove, in a SARP-network setting with investigators (Israel, Levy, Wechsler) most pharma severe-asthma programs already know, that EBC produces decision-grade analytical performance and pathway-level pharmacodynamic-class data when collection and assay work are executed correctly.
The three questions worth asking before you spend on it
EBC is a specialized tool, not a general one. Whether it belongs in a given early-phase program follows from three questions, all answerable before the protocol is written.
- Is the pharmacodynamic question lung-compartment specific? If blood markers, FeNO, or spirometry can answer it, EBC adds cost without unique value.
- Does the question require a noninvasive airway measure that FeNO cannot provide? If the PD question is Th2-driven eosinophilic inflammation, FeNO is the right tool. If it involves oxidative stress, neutrophilic inflammation, pH, cytokines, leukotrienes, or multi-pathway biology, EBC is the appropriate noninvasive choice.
- Is there a pre-specified analysis plan with a defined downstream use? Define what a meaningful signal looks like, how between-day variability will be managed, and what decision the data could influence. Without this, the sub-study won’t produce interpretable data.
A yes to all three is the signal that EBC belongs in the program. The worst outcome is collecting without thinking through these questions — which produces data that exists but cannot be acted on, the same evidentiary gap reproduced at cost.
The decision is made once
An EBC sub-study in a Phase II respiratory POC trial runs roughly $125K to $320K depending on panel size and site count — about 1–3% of a typical $5M–$20M Phase II budget. The right comparison isn’t sub-study cost against trial cost. It’s sub-study cost against the cost of the decision the sub-study informs. A go/no-go that sends a program into Phase III, or terminates one that could have been redesigned, is worth orders of magnitude more than $300K to get right.
EBC doesn’t make that decision. It provides the proximal biological context that prevents the decision from being made in the dark — and the data that informs Phase III endpoint selection, responder enrichment, and dose-PD modeling on directly measured airway biology rather than systemic proxies.
Every trial that closes without lung biology collected has answered the collection question by default. That default is not neutral. The option to collect exists once, at the design stage. The cost of getting it wrong becomes visible later — when the question cannot be reopened.
Evaluating EBC for your program
If you’re considering whether noninvasive airway pharmacodynamic measurement addresses a gap in your early-phase program, our team is available to discuss study design, collection standardization, and biomarker selection. The application note this post is drawn from — a longer technical framework with full bibliography and a worked Phase II COPD example — is available for download by clicking the button below.
Application Note Download