Ventilator-associated pneumonia biomarker evaluation (VIBE) study: protocol for a prospective, observational, case-cohort study

Abstract

Introduction

Current guideline-recommended antibiotic treatment durations for ventilator-associated pneumonia (VAP) are largely standardised, with limited consideration of individual patient characteristics, pathogens or clinical context. This one-size-fits-all approach risks both overtreatment—promoting antimicrobial resistance and adverse drug events—as well as undertreatment, increasing the likelihood of pneumonia recurrence and sepsis-related complications. There is a critical need for VAP-specific biomarkers to enable individualised treatment strategies. The Ventilator-associated pneumonia Biomarker Evaluation (VIBE) study aims to identify a dynamic alveolar biomarker signature associated with treatment response, with the goal of informing personalised antibiotic duration in future clinical trials.

Methods and analysis

VIBE is a prospective, observational, case-cohort study of 125 adult patients with VAP in Michigan Medicine University Hospital intensive care units. Study subjects will undergo non-bronchoscopic bronchoalveolar lavage on the day of VAP diagnosis (Day 1) and then on Days 3 and 5. Alveolar biomarkers (quantitative respiratory culture bioburden, alveolar neutrophil percentage and pathogen genomic load assessed via BioFire FilmArray polymerase chain reaction) will be assessed. An expert panel of intensivists, blinded to biomarker data, will adjudicate each patient’s Day 10 outcome as VAP clinical cure (control) or treatment failure (case). Absolute biomarker levels and mean-fold changes in biomarker levels will be compared between groups. Data will be used to derive a composite temporal alveolar biomarker signature predictive of VAP treatment failure.

Ethics and dissemination

Ethical approval was obtained from the University of Michigan Institutional Review Board (IRB #HUM00251780). Informed consent will be obtained from all study participants or their legally authorised representatives. Findings will be disseminated through peer-reviewed publications, conferences and feedback into clinical guidelines committees.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• Prospective study of longitudinal alveolar sampling in ventilator-associated pneumonia (VAP).

• Use of alveolar biomarkers already employed in pneumonia diagnostics to facilitate clinical translatability.

• Robust criteria for diagnosis of VAP and adjudication of clinical cure.

• Single-centre scope limits generalisability.

• Observational nature precludes definitive causal inference.

Introduction

Ventilator-associated pneumonia (VAP) is among the most common and morbid nosocomial infections in critically ill patients requiring invasive mechanical ventilation. VAP affects up to 40% of patients within intensive care units (ICUs) and is associated with increased mortality, ventilator-dependence, length of stay and healthcare expense.¹,⁷

Despite its clinical importance, treatment for VAP remains largely uniform. Current expert consensus treatment guidelines recommend a 7 day antibiotic treatment duration (ATD) for all VAP patients, regardless of patient or pathogen.^{8 9} This ‘one-size-fits-all’ treatment approach results in antibiotic overtreatment of many patients—predisposing them to complications including antimicrobial resistance and drug-related toxicity—while others are undertreated, facilitating pneumonia recurrence and sepsis-related organ dysfunction.¹⁰,²¹ VAP represents a complex interaction between diverse patients, pathogens and clinical conditions. Our inability to personalise VAP ATDs is inefficient and harmful.

Efforts to personalise treatment durations have been hindered by a lack of VAP-specific biomarkers to guide ATDs. Existing tools to guide antibiotic use—including normalisation of abnormal bedside clinical variables and systemic biomarkers such as procalcitonin or C-reactive protein—are not specific for pneumonia, often fail to normalise in a timely or reliable manner (even in patients who have clinically improved), and have failed to meaningfully impact pneumonia-directed antibiotic use in real-world scenarios.²²,²⁹ Existing biomarkers reflect systemic inflammation, not necessarily the host and pathogen response to pneumonia-directed antibiotic treatment, limiting their utility in guiding VAP ATDs.

Alveolar biomarkers, measured directly from bronchoalveolar lavage (BAL) fluid, offer a unique opportunity to evaluate local infection dynamics, assessing host and pathogen response to treatment directly at the site of disease. Although historically dependent on bronchoscopy, alveolar biomarker sampling can now be performed via non-bronchoscopic BAL (NB-BAL)—an inexpensive, safe and reproducible method for obtaining alveolar fluid in critically ill patients.⁶³⁰,³² To date, no prospective study has systematically evaluated the utility of alveolar biomarkers in predicting VAP treatment response. The Ventilator-associated pneumonia Biomarker Evaluation (VIBE) Study seeks to fill this critical knowledge gap by identifying robust, translatable alveolar biomarkers of VAP treatment failure to support personalised antimicrobial strategies.

Methods and analysis

Study design

The VIBE Study is a prospective, single-centre, observational case-cohort study conducted within ICUs at Michigan Medicine’s University Hospital and the Frankel Cardiovascular Centre, both part of a tertiary referral academic medical centre, which collectively comprise over 100 ICU beds. The study will enrol adult patients receiving invasive mechanical ventilation within participating ICUs diagnosed with VAP, defined by a positive BAL culture and fulfilment of rigorous clinical criteria. Patients will be enrolled within 1–3 calendar days of VAP diagnosis, provided they meet all inclusion criteria and none of the study’s key exclusions, which include contraindications to BAL, early extubation and anticipated death or hospice discharge. The study began enrolling patients in October 2024, is currently meeting prespecified enrolment quotas and is anticipated to complete participant accrual in Fall 2027.

Study participants will undergo NB-BAL sampling at three distinct time points during their VAP antimicrobial treatment course: the day of VAP diagnosis (Day 1), Day 3 following VAP diagnosis and Day 5 following VAP diagnosis. NB-BAL fluid will be assayed for quantitative respiratory culture bioburden, alveolar neutrophil percentage and pathogen genomic load assessed via BioFire FilmArray polymerase chain reaction (PCR). The primary study objective is to compare alveolar biomarker trajectories in patients who experience treatment failure (cases) versus those who experience clinical cure (controls) (figure 1).

Figure 1. Schematic model of alveolar biomarkers as predictors of VAP treatment failure. The ventilator-associated pneumonia biomarker evaluation study posits that alveolar biomarkers—including quantitative respiratory culture bioburden, bronchoalveolar lavage fluid neutrophilia and pathogen genomic load via BioFire FilmArray polymerase chain reaction—reflect persistent infection and can predict VAP treatment failure. VAP, ventilator-associated pneumonia.

Patient eligibility criteria and recruitment

VIBE Study eligibility criteria are shown in Box 1. Adult patients aged ≥18 years receiving invasive mechanical ventilation via endotracheal tube or tracheostomy within study ICUs will be enrolled if they meet clinical and microbiological criteria for VAP within 1–3 calendar days of a qualifying BAL procedure. Eligible patients must have a BAL culture positive for a pathogenic organism, accompanied by BAL cell count and differential performed on Day 1. Additional inclusion criteria require evidence of systemic and pulmonary inflammation consistent with a predefined VAP case definition. Key exclusions include contraindications to BAL (eg., recent airway surgery, thrombocytopenia, severe hypoxaemia, grossly bloody secretions), anticipated death or hospice discharge within 48 hours, early anticipated extubation, exclusive receipt of nocturnal ventilation and established pregnancy at enrolment.

Box 1. Ventilator-associated pneumonia biomarker evaluation study eligibility criteria.

Inclusion criteria

• Age ≥18 years at time of study enrolment.

• Hospitalisation in the Michigan Medicine study ICU at the time of enrollment.

• Use of invasive mechanical ventilation delivered via endotracheal intubation or tracheostomy tube at the time of index VAP diagnosis and at the time of enrollment.

• BAL respiratory culture with growth of pathogenic organism at any level of quantification and BAL cell count/differential performed on VAP Day 1.

• Fulfilment of study Pneumonia Case definition 1–3 calendar days prior to enrolment (see below).

Exclusion criteria

• Contraindication to BAL performance (major lung/airway surgery in last 30 days, gross blood in endotracheal secretions, platelet count <50 000 cells/mm3, PaO₂/FiO₂ ratio <80).

• Anticipated discharge to hospice/comfort care or death within two calendar days of enrollment.

• Extubation prior to or on Day 3 following index VAP diagnosis.

• Exclusive use of nocturnal invasive mechanical ventilation.

• Women with established active pregnancy previously diagnosed as part of routine clinical care.

Pneumonia case definition

• At least one of the following within +/−1 calendar day of respiratory collection:

  • New or worsening infiltrate on chest radiograph performed within 48 hours of index VAP.

  • Pre-existing diagnosis of acute respiratory distress syndrome.⁸

  • Increase in FiO2 ≥0.1 OR increase in PEEP ≥3 cm H₂O sustained for at least 6 hours (compared with minimum recorded value from prior calendar day).

  • Purulent endotracheal secretions.

  • For patients requiring ECMO, minimum flow increase >10% OR minimum sweep increase >20% over 24 hours (compared with minimum recorded value from prior calendar day).

• At least one of the following within +/−1 calendar day of respiratory collection:

  • Abnormal temperature (temperature ≥38.3°C or <36°C) within +/−1 calendar day of Index VAP.

  • Abnormal white cell count ≥10 x 10⁹ cells/L or <4 x 10⁹ cells/L within +/−1 calendar day of Index VAP.

  • A condition which could mask systemic response to infection:

    • Immunocompromised patient.

    • Patient requiring CRRT.

    • Patient requiring ECMO.

Eligible patients will be identified through real-time electronic medical record (EMR) alerts and active screening of ICU patient lists by study personnel. To facilitate identification of eligible participants, an EMR-embedded clinical decision support tool, featuring a best practice alert, will prompt ICU clinicians ordering respiratory cultures to perform NB-BAL for VAP evaluation when no contraindications exist (online supplemental materials). This guideline-concordant strategy represents the standard of care and has been successfully implemented in prior institutional initiatives.³³ Participation in the study will be financially incentivised using prepaid gift cards.

Informed consent

Informed consent will be obtained from all study patients or their legally authorised representatives (LARs). Given the critical illness of the study population, many patients will lack decision-making capacity at the time of enrolment. In such cases, informed consent will be obtained from a LAR, in accordance with Michigan state law and institutional policy. Participants who regain capacity during the study period will be re-consented. Consent forms and procedures have been reviewed and approved by the University of Michigan Institutional Review Board.

Study procedures and assessments

BAL performance

NB-BAL will be performed using a standardised institutional protocol by Respiratory Therapists. A sterile tracheobronchial catheter is inserted through the endotracheal tube or tracheostomy and advanced into the distal airways, after which 30 mL’s of sterile saline is instilled and aspirated. Samples will be immediately transported following collection for laboratory analysis.

Study participants will undergo NB-BAL on Days 1, 3 and 5 of antimicrobial therapy (with VAP Day 1 corresponding to the calendar date of index BAL culture performance). Before each research BAL, patients will undergo a safety assessment to confirm the absence of procedural contraindications (occurrence of major lung/airway surgery within 30 days, gross blood in endotracheal secretions, platelet count <50 x 10⁹ cells/L, PaO₂/FiO₂ ratio <80).

Day 1 NB-BAL specimens will be obtained as part of routine clinical care, with excess biospecimen reclaimed from the Clinical Laboratory for subsequent research testing. Days 3 and 5 NB-BAL research specimens will be collected by respiratory therapists. If a clinical team plans performance of BAL or NB-BAL for clinical indications on Study Days 3 or 5, the clinically obtained BAL fluid will be used for all research testing to avoid redundant procedural interventions. BAL sampling is performed per institutional policy in patients with endotracheal tubes or tracheostomies. Patients liberated from mechanical ventilation prior to the study Day 5 assessment will not undergo NB-BAL testing.

BAL fluid laboratory testing

NB-BAL fluid from Days 1, 3 and 5 will undergo the following testing:

1. Bacterial bioburden of pathogenic organisms (expressed as colony forming units per millilitre (CFU/mL)) will be obtained from quantitative BAL cultures performed according to the Michigan Medicine Clinical Microbiology Laboratory’s standard operating procedures supplemented by secondary review of culture plate images from the BD Kiestra Total Lab Automation System (Beckton, Dickinson and Company; Sparks, MD) for confirmation of pathogen quantitation. Bacterial identification will be performed by matrix-assisted laser desorption ionisation-time of flight mass spectrometry using the Bruker Biotyper system (Bruker Daltonics, Billerica, MA).

2. The percentage of polymorphonuclear cells among total nucleated cells will be obtained from BAL cell count/differential testing (using the Sysmex 9100 Automated Haematology System).

3. Genomic load of pathogenic organisms (expressed as semi-quantitative genomic loads (copies/mL)) will be obtained from the Biofire FilmArray Pneumonia panel (using the Biofire FilmArray Torch system).

Residual research BAL fluid will be processed and stored for future biomarker analyses. Of note, clinical providers will be blinded to the results of the research NB-BAL testing.

Data collection

Demographic and clinical data will be collected at baseline and will include age, sex, race/ethnicity, body mass index, medical comorbidities, reason for ICU admission and mechanical ventilation specifications. Clinical data collected at Days 1, 3, 5, 10 and 28 will include vital signs, laboratory values, ventilator settings, medication usage (antibiotics, steroids, vasopressors), radiographic findings, microbiologic culture results and Sequential Organ Failure Assessment (SOFA scores). Data will be abstracted from the EMR by study personnel using validated case report forms.

Adverse events

The VIBE study defines an adverse event (AE) as any untoward medical occurrence—sign, symptom or disease—regardless of its relationship to study procedures. Serious adverse events (SAEs) are those causing death, life-threatening illness, prolonged hospitalisation, persistent disability or those requiring intervention to prevent these outcomes. Because all participants are critically ill and many complications are expected as part of their illness trajectory, only AEs within the known procedural risk profile NB-BAL will be systematically captured. An expert clinician will judge each event’s relatedness to the NB-BAL, classifying causality from ‘definitely related’ to ‘not related’ based on timing and clinical context.

All AEs are recorded on case report forms with details on onset, severity, relationship and resolution. Reportable events (SAEs or unexpected AEs deemed related or possibly related to NB-BAL) must be notified to the principal investigator within two calendar days and to the Institutional Review Board (IRB) and independent safety monitor within 24 hours, with a formal written report to the IRB within 15 days. AE assessments occur 1 day after each BAL procedure (Days 3 and 5), drawing on clinical data reviews.

Case definitions and biomarker endpoints

Case definitions

The VIBE study seeks to evaluate associations between alveolar biomarker signatures and clinical recovery from VAP. However, VAP represents a heterogeneous clinical syndrome without a universally accepted gold standard for defining treatment response or validating biomarkers. Outcome definitions used in prior clinical trials often fail to capture important gradations in therapeutic response and are limited by methodological challenges—including competing risks, conflation of outcomes with differing clinical significance and reliance on subjective adjudication of clinical cure.³⁴,³⁶ In critically ill patients with VAP, distinguishing improvements attributable to pneumonia resolution is particularly challenging due to overlapping causes of respiratory failure and variable illness trajectories.

To address this complexity, the VIBE study will assess biomarker performance using a primary case definition (Day 10 Clinical Cure), as well as multiple exploratory definitions of pneumonia recovery (28-day ventilator-free survival, 28-day recurrent pneumonia-free survival). This multimodal approach aligns with precedents in acute respiratory distress syndrome (ARDS) and sepsis research, where investigators have employed multiple outcome definitions—such as ventilator-free days, organ failure scores and survival endpoints—to validate biologic phenotypes and biomarkers in the absence of a single definitive measure of recovery.³⁷,⁴⁰

The primary case definition for VAP treatment failure in the VIBE study is failure to achieve clinical cure by Day 10 following VAP diagnosis. Clinical cure will be adjudicated by an expert panel of ICU Medicine and Infectious Diseases physicians using standardised criteria (Box 2). The study definition of clinical cure is adapted from a previously validated adjudication protocol for severe pneumonia.⁴¹ It incorporates evidence of improvement or resolution of clinical signs and symptoms (eg, fever, leukocytosis, purulent secretions, radiographic findings), absence of new or escalating antibiotic therapy and no evidence of pneumonia recurrence following antibiotic cessation. All adjudicators will be blinded to the research NB-BAL results. Each case will be reviewed independently by two adjudicators, with a third reviewer resolving any discrepancies.

Box 2. Day 10 clinical cure definition.

Criteria

• Patient alive and not receiving/designated to receive end-of-life/comfort care on Day 10.

• Pneumonia-directed antibiotics stopped without recurrent pneumonia within 48 hours of discontinuation OR pneumonia-directed antibiotics continued but signs/symptoms of pneumonia (white cell count, endotracheal secretions, oxygenation, chest radiography) are stable/improved.

• Causative pathogen not isolated from further respiratory specimens obtained as part of routine clinical care in a context indicating treatment failure (perceived clinical indication for repeat culture, timing of repeat respiratory culture relative to index VAP and Gram stain results may influence decision).

• Clinical manifestations of index VAP (fever, secretions, white cell count, hypoxemia, septic shock) are improved/resolved (abnormal clinical signs in critically ill patients may be multifactorial; persistent abnormal clinical signs attributable to a convincing non-VAP condition do not necessarily indicate treatment failure).

• Patient able to wean from mechanical ventilation, initiate spontaneous breathing trials or return to pre-VAP ventilator/ECMO settings.

Secondary case definitions include: (1) 28-day ventilator-free survival, defined as being alive and free from invasive mechanical ventilation on Day 28 and (2) 28-day recurrent pneumonia-free survival, defined as being alive on Day 28 without evidence of pneumonia recurrence. Recurrent pneumonia is defined for study purposes as growth of the index VAP pathogen from a respiratory culture obtained ≥48 hours after discontinuation of pneumonia-directed antibiotics and fulfilment of the study’s pneumonia case definition (see Box 1). These secondary outcome constructs have been employed in multiple prior VAP trials and provide complementary information on recovery and relapse patterns relevant to biomarker performance.⁴²,⁴⁴

Biomarker endpoints

Candidate biomarkers for the VIBE study were selected based on preliminary evidence supporting their relevance in VAP and their current use in clinical practice for diagnostic purposes.^{6 32} Although these biomarkers have not yet been operationalised to assess treatment response, their established use in existing clinical care enhances the potential for rapid clinical translation if shown to be prognostically informative.

Primary and secondary study endpoints are shown in table 1. The VIBE Study will assess both mean absolute alveolar biomarker levels—averaged across Days 1, 3 and 5 following index VAP diagnosis—and mean-fold declines in study patients between Day 1 and the average of Days 3 and 5. Pathogen bioburden quantified by culture will serve as the primary endpoint. Secondary endpoints include alveolar neutrophil percentage and pathogen genomic load (copies/mL). Exploratory analyses will incorporate semi-quantitative Gram stain assessments of bacterial burden.

Table 1. Primary and secondary study biomarker endpoints.

Endpoint category	Biomarker	Metric
Primary	Pathogen bioburden (CFU/mL)	Mean absolute (average of Days 1, 3 and 5)
		Mean-fold decline (Day 1 vs average of Days 3 and 5)
Secondary	Alveolar neutrophilia (% polymorphonuclear leukocytes on BAL differential)	Mean absolute (average of Days 1, 3 and 5)
		Mean-fold decline (Day 1 vs average of Days 3 and 5)
Secondary	Bacterial genome copies/mL (BioFire FilmArray Pneumonia Panel)	Mean absolute (average of Days 1, 3 and 5)
		Mean-fold decline (Day 1 vs average of Days 3 and 5)

BAL, bronchoalveolar lavage; CFU, colony forming unit.

For patients with multiple organisms on index VAP cultures, the VIBE Study will classify causative pathogens based on pathogenicity and quantitative burden: non-pathogens or low-burden organisms (<10³ CFU/mL) will be considered contaminants, while high-burden pathogens (>10³ CFU/mL) will be considered causative. In cases of polymicrobial infection, biomarker values will be averaged across pathogens for analysis.

Statistical analysis plan

Descriptive analyses

We will compare baseline characteristics between patients who achieve clinical cure (controls) and those who experience treatment failure (cases) by Day 10. These comparisons will include demographic characteristics, comorbid conditions, markers of illness severity, prior healthcare exposures and microbiological aetiology of pneumonia. Standard hypothesis testing will be used: χ² or Fisher’s exact tests for categorical variables, and t-tests or Wilcoxon rank-sum tests for continuous variables, depending on distributional assumptions.

Biomarker analyses

Alveolar biomarker data will be expressed as CFU/mL (quantitative culture), percent neutrophils (BAL differential) and genome copies/mL (BioFire FilmArray), with transformations applied as appropriate to normalise distributions. Quantitative culture results and BAL neutrophilia will be analysed as continuous variables; multiplex PCR results, reported semi-quantitatively, will be treated as ordinal variables.

We will first compare both the absolute levels and response trajectories of each biomarker between cases and controls. Crude (unadjusted) comparisons will evaluate: (1) Average biomarker values across Days 1, 3 and 5, and (2) Mean fold-change from Day 1 to the average of Days 3 and 5. For adjusted analyses, we will fit mixed-effects models to account for within-subject correlation across timepoints. Linear mixed-effects models will be used for continuous biomarker values, and ordinal models for semi-quantitative PCR results. Case/control status will be included as a fixed effect, and random intercepts will account for within-subject clustering.

We will include interaction terms between group and time to assess differential biomarker trajectories. We will also explore microbiological aetiology (eg, Staphylococcus aureus, non-fermenting Gram-negative bacilli or Enterobacterales) as potential effect modifiers within this modelling framework. Of note, all biomarker analyses will be applied to both the primary outcome (Day 10 clinical cure) and secondary outcome constructs (eg, Day-28 ventilator-free survival and recurrent pneumonia-free survival).

Missing data and sensitivity analyses

Missing data will be examined to assess the pattern and mechanism of missingness. For variables with missing clinical data, we will use multiple imputation by chained equations. For patients with missing Day 5 alveolar biomarker data—anticipated primarily due to early death or extubation precluding bronchoscopy—we will examine biomarker profiles within both an intention-to-treat framework (all enrolled study subjects), as well as a per-protocol framework (only patients with Days 3 and 5 research BALs completed). We will additionally apply inverse probability of censoring weights to account for informative missingness.

Composite biomarker signature modelling

To identify temporally-informed alveolar biomarker signatures predictive of treatment failure, we will develop multivariable models incorporating both binary thresholds of biomarker levels at each time point and derived change scores of biomarker levels over time. We will evaluate model discrimination using the area under the receiver operating characteristic curves and assess fit using standard diagnostics. Model performance will be compared and validated against a reference clinical model that includes established predictors of VAP treatment failure.

Sample size justification

Based on retrospective data and assumptions of biomarker distributions, a sample size of 125 patients provides 80% power (α=0.05) to detect both (1) A 0.91 log difference in quantitative culture values (SD=1.7) and (2) A 6.7% absolute difference in alveolar neutrophilia (SD=12%). Assuming a 2:1 control-to-case ratio and accounting for 10%–15% attrition due to early death or extubation, this sample size is sufficient for primary and key secondary analyses.

Quality control and data management

An independent safety officer will review all serious AEs. Study quality assurance procedures include regular team meetings, quarterly reporting to the ISO on study performance, and biannual source data validation with full case review if discrepancies exceed 10%. Data will be collected and managed using REDCap, which incorporates built-in quality checks. Study records will be retained for a minimum of 10 years.

Patient and public involvement

None.

Ethics and dissemination

The study protocol was approved by the University of Michigan IRB (HUM00251780). The study will comply with the Declaration of Helsinki and Good Clinical Practice guidelines. Informed consent or LAR consent will be obtained from all participants. Study results will be submitted for peer-reviewed publication and presented at national and international critical care and infectious diseases conferences, and findings will also be disseminated via feedback into clinical guidelines committees. Data will be shared in accordance with NIH data sharing policies.