Citations

Abstract Rationale Oscillometry measures the mechanical properties of the respiratory system using the response to pressure signals superimposed on tidal breathing. The JaegerTM IOS device generates “pulse packets” of harmonic frequencies between 5 and 35 Hz in 5 Hz intervals, while the Resmon ProTM V3 FOT and Thorasys TremofloTM C-100 employ prime pseudorandom sine waveforms between 5-19 Hz and 5-37 Hz, respectively. Prior studies have raised concerns about harmonic distortion with the IOS technique, but a direct comparison of reproducibility using current devices has not been performed. Methods Healthy volunteers with normal spirometry performed weekly oscillometry following ERS technical standards using JaegerTM and ResmonTM or ThorasysTM oscillometers in random order for 10 consecutive weeks. Z scores were calculated using Berger and Oostveen normative equations. Means, pooled within-subject distributions and interquartile range, standard error (SEM), coefficient of variation (CVw%), absolute and relative minimal detectable change (MDC, MDC%), and between device and subject intraclass correlation coefficients (ICC) were calculated for R5, R10/R11, R20/R19, R5-R20/R5-R19, X5, AX, and Fres and compared. Results Fifteen participants completed 10 weekly measurement sessions with JaegerTM and ResmonTM and twelve with JaegerTM and ThorasysTM. One subject did not complete the ResmonTM analysis due to intolerance; another subject only completed three weeks of ThorasysTM prior to work assignment change. The majority (65%) were female with mean (SD) age of 37.7 (10.2) years and BMI 27 (5) kg/m2. Rrs reproducibility was excellent across devices (ICC > 0.89), with more variability for X5, AX, and Fres (ICC 0.66 - 0.93) and little difference between devices (Figure 1). Small, systematic increases in mean Rrs were seen with JaegerTM compared to ResmonTM (5-10%, p < 0.008) and TremofloTM (5-12%, p < 0.02), with higher mean AX vs ResmonTM (44%, p = 0.002) and lower Fres vs TremofloTM (19%, p = 0.02). There was high ICC (0.97-0.99) across ResmonTM and TremofloTM Rrs comparisons with minimal bias. Intra-subject reproducibility was similar between devices. MDC% was 22.7-39.4% for Rrs values, 52.7-70.8% for X5, and 131-142% for AX. Conclusions JaegerTM, ResmonTM, and ThorasysTM oscillometers performed similarly in healthy adults over 10 weeks, with small but systematic measurement differences seen with the JaegerTM device. Rrs and Fres values demonstrated the greatest reproducibility both within and between subjects. Reactance measurements had less between-device agreement and higher MDC. Rrs, X5, and AX changes of up to 40%, 70%, and 140% over time may be due to normal subject and measurement variation. This abstract is funded by: 2024 Pulmonary, Critical Care, and Sleep Medicine Schuen Family Innovation Award

Abstract Rationale The governing mechanisms of lung regeneration are still poorly understood. Studies have observed the lung’s regenerative capacity using the pneumonectomy (PNX) model, where removal of the left lung stimulates compensatory growth and neoalveolarization. However, this process is inhibited if a physical object is used to replace the extracted tissue, demonstrating that the lung can sense the available space and regenerate in response. The mechanism of this mechanosensing is unknown. As outermost tissue layer of the lung, the pulmonary mesothelium is well situated to sense the surrounding environment. We and others have hypothesized that the mesothelium can become activated in cases of injury or regeneration, but this has not been fully explored. Understanding the mechanisms controlling neoalveolarizaiton may be key to unlocking the lung’s natural regenerative capacity in humans. Methods & Results In this study, we examined the location and fate of the mesothelium during regeneration using fluorescent lineage tracing and the PNX model. Specifically, we used the mesothelial-specific Wt1CreERT2 crossed with a lox-stop-lox-RFP reporter (Wt1(CreERT2/+);LSL-RFP(+/-)). Following PNX, we collected the lineage-labeled cells along with the CD326+ and CD326-/CD45-/CD31- populations and performed scRNA-sequencing. As quantified by flow-cytometry, the number of RFP+ cells increased following PNX, indicating proliferation of the mesothelium. Clustering analysis elucidated the emergence of 3 activated clusters of mesothelium following PNX, representing 12.9% of the RFP+ population in Sham and 66% in PNX. These activated cells exhibited decreased expression of mesothelial hallmarks and increased expression of extracellular matrix and epithelial-to-mesenchymal transition genes. Using immunofluorescence and confocal microscopy, we observed lineage-labeled mesothelium deep within the parenchyma after PNX. We observed none or singular invaded mesothelium in no-procedure control animals and only minimal occurrence in animals exposed to Sham surgery. Compared to Sham controls, PNX post-surgical (D)ay 3, D7, and D14 had 8.5, 32.7, and 9.2-fold more cells, respectively, in the parenchyma. Moreover, in aged (18 month) mice, where regeneration is known to be significantly reduced, we likewise observed minimal mesothelial invasion following PNX. These cells were often found near or directly wrapping around progenitors of the alveolar epithelium, AT2 cells, whose proliferation is essential to neoalveolarization. Using time-lapse microscopy of precision cut lung slices, we observed the invading mesothelial cells to be highly dynamic, sending out multiple transient protrusions and interacting with neighboring cells. Conclusion Together these data support strong and potentially functional interactions between activated and invading mesothelial cells and lung regeneration following PNX. This abstract is funded by: NIH

Abstract Rationale Normal lung aging is characterized by progressive structural and functional decline that reduces respiratory efficiency. Aging is associated with excessive reactive oxygen species production and impaired homeostatic pathways that maintain mitochondrial quality control, involving mitophagy (removal of damaged mitochondria) and mitochondrial biogenesis (generation of new mitochondria). In human airway smooth muscle (hASM), reduced mitochondrial quality control can lead to the accumulation of damaged mitochondria, compromising ATP production and decreasing cellular viability. Maintaining healthy mitochondria involves a balance between mitophagy and mitochondrial biogenesis. A key regulator of this balance is PARIS, normally a target of the PINK1/pParkinS65 mitophagy pathway. PARIS transcriptionally inhibits PGC1α and NRF2 expression, key regulators of mitochondrial biogenesis and redox homeostasis. Furthermore, NRF2 can upregulate PINK1 expression, enhancing mitophagy. Therefore, we hypothesize that aging involves unmitigated mitochondrial dysfunction due to accumulation of PARIS and an imbalance between the removal of damaged mitochondria (mitophagy) and mitochondrial biogenesis. Methods hASM cells were dissociated from bronchiolar tissue samples from patients with no history of smoking or respiratory diseases and cultured (≤3 passages). Prior to experimentation, hASM cells from young (≤40 yrs) and old (≥70 yrs) patients (female and male) were serum deprived and phenotyped. In both age groups, targeted siRNA-mediated knockdown of PARIS was also performed. The extent of mitochondrial damage in hASM cells was determined by changes in mitochondrial membrane potential (Ψm) using TMRM (250 nM), and the extent of co-labeling of MitoTracker Red (200 nM) and CellLight GFP. Mitochondrial turnover was quantified by confocal imaging of hASM cells transfected with pMitoTimer. Mitophagy was determined by PINK1 recruitment to mitochondria, pParkinS65 and ubiquitin (pUbS65) phosphorylation. Total PARIS and pParkinS65-ubiquitinated PARIS (by immunoprecipitation) levels were determined. Mitochondrial biogenesis was determined by PGC1α and NRF2 mRNA and protein expression and mitochondrial DNA (mtDNA) copy number. Results In hASM cells from older patients, there was extensive mitochondrial damage, decreased PINK1 recruitment to mitochondria and decreased pParkinS65 and pUbS65 phosphorylation. Total PARIS accumulation increased due to decreased PARIS ubiquitination. Also, PGC1α and NRF2 expression, and mtDNA copy number decreased, likely due to PARIS-mediated inhibition. Knockdown of PARIS increased PGC1α and NRF2 expression, thereby promoting mitochondrial biogenesis and NRF2-mediated antioxidant responses and increasing PINK1 expression, thereby activating mitophagy. Conclusion Aging results in impaired mitochondrial quality control in hASM cells. The PARIS-NRF2 axis presents a promising target for therapeutic intervention to mitigate oxidative damage, maintain mitochondrial homeostasis and improve lung function with aging. This abstract is funded by: HL157984 (GCS) and AG44615 (GCS)

Abstract Rationale Pulmonary Fibrosis is characterized by excessive accumulation of extracellular matrix (ECM), primarily collagen I, leading to progressive lung scarring and loss of function. Many anti-fibrotic compounds have been found to attenuate collagen deposition. However, identifying compounds that stimulate collagen I degradation, a necessary step in fibrosis clearance, is less common. To identify such compounds, we developed in vitro and ex vivo assays to quantify extracellular collagen I and collagen I degradation products. Methods Primary lung fibroblasts (passage 1-3, N = 5) from patients with idiopathic pulmonary fibrosis (IPF) were cultured in DMEM/F12 with 2% FBS, 2 ng/mL TGF-β, and 50 µg/mL ascorbic acid for 72 hours to generate collagen-rich cell-derived ECM. Cells were then treated with test compounds for 24 hours in the same media. Nuclear labeling with Hoechst 33342 was used to quantify cell numbers by automated microscopy. Decellularization was then performed with a solution containing 20 mM ammonium hydroxide and 0.5% Triton X-100. Cell-derived ECM was then fixed, stained with anti-Collagen I antibody, and quantified by automated microscopy. Collagen I degradation was further assessed by measuring type I collagen C-telopeptide (ICTP) via ELISA in cell culture media. To investigate the role of matrix metalloproteinase-13 (MMP-13) in collagen degradation, we used T26-C, a selective MMP-13 inhibitor, or transfected cells with MMP-13 siRNA. Ex vivo validation of collagen-degrading activity was conducted by measuring type I collagen N-terminal telopeptide (NTX-I) in the culture media of human precision-cut lung slices (PCLS, N = 3) generated from upper and lower lobes of IPF tissue and stimulated with select compounds. Results Thirty-four candidate compounds (metabolic, epigenetic, and cAMP modulators) were tested for their ability to reduce collagen I levels in cell-derived ECM. Six hits and two positive controls were then validated using ICTP ELISAs. In total five compounds were found to stimulate collagen degradation: EGCG (1 μM), Nintedanib (0.25 μM), SB431542 (1 μM), and two dopamine receptor agonists, A68930 (10 μM) and Dihydrexidine (10 µM). MMP-13 inhibition or knockdown prevented collagen degradation by all five compounds. In human IPF PCLS, collagen degradation was confirmed in response to EGCG, Nintedanib and SB431542 but not Dihydrexidine, with much greater responses seen in PCLS from upper lobes than lower lobes. Conclusion Collagen degradation can be stimulated in cultured fibroblasts and IPF PCLS. These systems will assist in uncovering mechanisms and regulators of collagen degradation and may pave the way to better fibrosis-resolving therapies. This abstract is funded by: NIH U01 HL152967

Abstract Rationale Respiratory disease is a major cause of mortality in Down syndrome (DS; trisomy 21, T21). DS is considered a progeroid syndrome, with elevated cellular senescence (SEN) and oxidative stress. Because SEN shapes organogenesis, its dysregulation may impair lung morphogenesis. We previously found SEN/SASP (senescence-associated secretory phenotype) abnormalities in prenatal human T21 lungs. Here, we test whether these alterations persist and change after birth. Methods We analyzed human lung single-nuclear RNA (snRNA-seq) datasets (prenatal and postnatal) to quantify epithelial and mesenchymal SEN/SASP signatures. Sex- and age-matched T21 and non-T21 prenatal and postnatal lungs were studied; fibroblasts were isolated as prenatal human lung fibroblasts (HFLF) and postnatal human lung fibroblasts (HLF). SEN/SASP were measured by RT-qPCR and immunofluorescence (IF) in tissue and HFLF/HLF. HFLF/HLF were assayed for SA-β-gal, oxidative stress (CellROX/MitoSOX), and Ca²+ dynamics (Fura-2 AM). 48h treatments with tunicamycin (1µM, SEN inducer), 4-phenylbutyrate (4-PBA, 2 mM, ER Stress inhibitor), or pyrroloquinoline quinone (PQQ, 20 μM, antioxidant) were performed on T21 and non-T21 HFLF/HLF. Conditioned media (CM) were collected after 48h from untreated cultures and then applied to cells of the opposite genotype for 24, 48, and 72h. Results Overall SEN profile was similar in epithelium and mesenchyme across ages, but pathway signatures differed. Epithelium shows an inflammatory SASP, while mesenchyme shows a less-inflammatory, remodeling/growth-factor SASP. These lineage patterns were more evident postnatally. Whole-tissue IF showed increased γ-H2AX and p21 in T21 vs non-T21 both prenatally and postnatally (p < 0.05). At baseline, T21 HFLF/HLF had higher CDKN1A, CDKN2A, TP53, CDKN2B, and SERPINE1 (p < 0.05; n = 5-6), with increased γ-H2AX/p21 IF (p < 0.01; n = 6), SA-β-gal (p < 0.05; n = 5-6), and ER/mitochondrial stress (p < 0.05; n = 5-6). Fura-2 AM showed greater Ca²+ release in T21 HFLF/HLF at baseline and after acetylcholine (10 μM, p < 0.01; n = 5). Mitochondrial-morphology genes (FIS1, DNM1L, MFN1) were elevated in T21 HFLF/HLF (p < 0.05; n = 6). In T21 HFLF, 4-PBA and PQQ each reduced SA-β-gal (p < 0.01; n = 6). In postnatal HLF, PQQ had the stronger effect (p < 0.05; n = 5). T21 CM induced progressively greater SEN/SASP over 24/48/72h in non-T21 cells (p < 0.05; n = 5-6), while non-T21 CM partially rescued T21-HLF phenotypes at 24h only (p < 0.05; n = 5-6). Conclusion These data indicate that lung senescence in T21 initiates in utero and intensifies after birth and reflects lineage-specific pathways. ER-stress-related mechanisms are more prominent prenatally, while mitochondrial/redox mechanisms are predominant postnatally. This evidence suggests age and lineage-dependent drivers of senescence that may help explain DS lung disease and highlight potential new mechanistic targets. This abstract is funded by: NIH/NHLBI

Abstract Introduction The inhalation of aerosolized therapeutics is a promising strategy for the targeted delivery of drugs and biologics to the lungs. We have previously demonstrated the efficacy and feasibility of nebulized exosomes using a platelet-derived exosome product (PEP: Rion, Inc.), which significantly attenuated emphysema in a murine model (PMID: 38325750). However, the PEP aerosol’s physical properties—such as size, tissue deposition, anatomic location, and stability remain incompletely characterized. To address this, we utilized a cascade impactor to define aerosol particle size and deposition patterns and employed immunofluorescence staining to assess the regional and anatomic localization of nebulized PEP within the murine lung. Methods PEP particle size and tissue deposition were determined by passing the aerosol through a Marple cascade impactor (CI, DSI), which was connected to an inhalation tower (DSI). PEP aerosols were generated by a vibrating mesh nebulizer (Aerogen, Cat No AG-AL1100) at a controlled velocity and humidity and delivered to the CI at a flow rate of 2L/min. Aerosols were separated by size and allowed to deposit at the corresponding stages as they passed through the pinholes of the stage, by design. The stages were weighed after nebulization, and the percentage of aerosol deposition was calculated to simulate the human respiratory tree, according to the manufacturer’s specifications. In parallel, fluorescence DiR-labeled PEP was nebulized to spontaneously breathing mice, and confocal microscopy was used to determine intrapulmonary distribution. Results A total of 2.25 mL of PEP reconstituted in 0.9% normal saline (11.4 mg/mL final concentration) was nebulized for 35 minutes at a flow rate of 2 L/min, yielding 3.5 mg of total aerosol deposition, which is 13.6% of the total PEP mass (25.65mg) being nebulized, across the CI stages. Of this, 74% of the aerosolized particles with a particle size ≤ 1µm and deposited in Stages 7, 8 & F, corresponding to the deep regions of the lungs, specifically regions beyond the alveolar ducts and within the alveoli. In vivo 3D confocal imaging corroborated these findings, revealing abundant DiR-labeled PEP deep into the alveolar space within the peripheral regions of all lung lobes. Conclusion The study result demonstrates the feasibility and capability of delivering aerosolized PEP deeply into the alveolar regions of the lung at the given pressure, humidity, flow rate and nebulizer settings. This study establishes a methodological framework for characterizing and optimizing aerosolized exosome therapeutics for precise, region-specific delivery within the respiratory tract. This abstract is funded by: Department of Defense, Mayo Clinic Funding

Abstract Rationale Premature infants exposed to supplemental oxygen are at increased risk of developing airway diseases such as bronchopulmonary dysplasia (BPD) and asthma, making it important to understand how oxygen detrimentally impacts developing airways. Even in moderation, hyperoxia promotes bronchial airway hyperresponsiveness (AHR) through effects on fetal airway smooth muscle (fASM), a cell type involved in impaired bronchodilation and remodeling. Iron dysregulation has been implicated in adult airway diseases such as asthma, COPD, and hyperoxic lung injury (HLI). Immature antioxidant systems in neonates and their increased susceptibility to iron-induced oxidative stress may lead to ferroptosis, contributing to pathologies observed in subsequent airway disease. The role of impaired iron metabolism, mediated by hyperoxia, remains largely unexplored in fetal airways. This study investigated the effects of increasing oxygen levels on antioxidant systems, iron metabolism, and lipid peroxidation in human fASM. Methods fASM from 18-20 week gestational age (canalicular stage; Mayo IRB exempt) were isolated and exposed to normoxia (21% O2) or increasing levels of hyperoxia (50%, 70%, and 90% O2) and treated with iron chelator, deferoxamine (DFO) [100uM] or reactive oxygen species (ROS) scavenger ferrostatin (Fer-1) [10uM] for 48h. Levels of intracellular iron (Fe2+) were determined using ferroOrange. Cell lysates were examined for protein expression of antioxidant markers (GPX4, SLC7A11, CBS, NRF2, KEAP1, and FSP1) and iron metabolism markers (TfR1, FTH1, and NCOA4). 4-HNE (lipid radical) expression was determined by immunofluorescence staining. Results Hyperoxia decreased SLC7A11, CBS, NRF2, KEAP1 (redox regulators), and iron storage (FTH1) at all concentrations. Iron influx (TfR1) increased under 50% and 70% O2 but not 90% O2. Cytosolic iron levels increased under hyperoxia in a dose-dependent manner and decreased with iron chelator, DFO. DFO further decreased FTH1 expression under 90% O2. 4-HNE (lipid radical) increased under 90% O2 and was reduced by ROS scavenger, Fer-1. Conclusion Our findings suggest that hyperoxia simultaneously impairs antioxidant systems and iron metabolism in fASM. Specifically, hyperoxia dose-dependently increases the labile iron pool which exhausts anti-ferroptotic mechanisms: effects which can be reduced by DFO and Fer-1. This implies the presence of a redox threshold tightly regulating iron accumulation in fASM. Overall, antioxidant impairment and iron metabolism dysfunction synergistically can contribute to pediatric airway disease: mechanisms which allow for the development of targetable strategies to blunt the impact of supplemental oxygen in the premature lung. This abstract is funded by: National Institutes of Health

Abstract Rationale Asthma in the elderly (AIE) is characterized by hyperreactive and thickened, fibrotic airways (remodeling) that are not adequately addressed by current therapies. Thus, understanding mechanisms that drive AIE becomes important. Here, airway smooth muscle (ASM) cells play important roles in normal airway physiology and asthma pathophysiology. Recent discoveries have highlighted the significance of mechanosensitive Piezo (Pz)-Piezo1 and Piezo2 in the lung. While Pz channels are highly expressed, their specific contributions to ASM function remains largely unexplored. Moreover, Pz channels play essential roles in regulating the intracellular calcium homeostasis, which further drives mitochondrial dynamics. Whether Pz signaling influences mitochondrial structure and function-a key determinant of ASM energetics and contractility is unknown and therefore of significant interest. In the present study, we hypothesize that Pz1 channel activation regulates mitochondrial structure and function via an increase in intracellular calcium in ASM, leading to metabolic changes and remodeling, characteristic of AIE. Methods Primary human ASM cells were isolated from lungs of young (<45y) and AIE (≥65y) individuals obtained post-surgery (Mayo Clinic IRB approved). Cells were exposed to mechanical stretch (mimicking breathing) in the presence of either Yoda1/Jedi2 (agonist) or GsMTx4 (inhibitor) for 24 hrs. Mitochondrial fractionation was performed, and protein expression evaluated for mitochondrial dynamics targets (MFN1, MFN2, OPA1, DRP1 and FIS1) and transcriptional markers (mt-Cox1, mt-Cox2, ATP5b and TFAM). The colocalization of mitochondria and Pz channels was observed by immunofluorescence staining. Mitochondrial morphology was examined with MitoTracker staining while function was assessed using Seahorse XF analysis. Intracellular calcium ([Ca2+]i) and mitochondrial calcium ([Ca2+]m) were measured using 5µM Fura-2/AM and 1µM Rhod-2 AM respectively with fluorescence microscopy. Results Confocal microscopy revealed that Piezo1 localizes to mitochondria in ASM. Pz1 activation promoted mitochondrial fission, evidenced by an upregulation of DRP1/FIS1 and downregulation of MFN1/OPA1, confirming a fission-dominant phenotype. Pz1 activation also increased mitochondrial branching while reducing branch length, indicating enhanced fragmentation, particularly in AIE ASM. Yoda1 treatment reduced mitochondrial oxygen consumption rates, suggesting a lower energy demand. Furthermore, Yoda1 increased both [Ca2+]i and [Ca2+]m responses to histamine (10µM), with a more pronounced effect in AIE ASM. Conclusion Overall, our findings suggest that mechanotransduction through Pz channels, particularly Pz1 regulates intracellular calcium regulation and mitochondrial structure and function in ASM, linking extracellular mechanical forces to intracellular energy homeostasis. This mechanism may contribute to the altered energetics and remodeling observed in AIE airway pathophysiology. This abstract is funded by: 24POST1241979 (NAB), R01- HL056470 (YSP), R01-HL177837 (CMP)

Abstract Rationale Extracorporeal membrane oxygenation (ECMO) is an increasingly utilized form of life support for patients with cardiac and/or respiratory failure. Fungal infections remain a serious challenge in patients supported by ECMO, with an estimated prevalence of up to 10% and high morbidity and mortality. While routine antifungal prophylaxis is uncommon in ECMO centers and widely debated, its impact on incidence, pathogenic organisms, resistance patterns, and outcomes has not been described. Methods This was a ten-year experience study of fungemia in patients receiving ECMO where most patients received antifungal prophylaxis. Institutional protocol was to prescribe prophylactic fluconazole based on risk stratification, substituted with echinocandins if indicated. Adults (18 years and older) with ECMO runs between 2014-2024 were screened for fungal organism growth on culture, then narrowed to those with positive blood cultures drawn in the window from 24 hours after cannulation through 72 hours after decannulation. Clinical, microbiological, and outcomes data were descriptively summarized. Institutional review board approved with a waiver of informed consent. Results Among 1,039 adult ECMO runs (979 patients), thirteen patients with fungemia were identified for a calculated incidence rate of 1.29 per 1000 ECMO days (95% CI 0.75 to 2.23). Twelve patients (92%) received antifungal prophylaxis; 85% had echinocandin and 38% azole exposure, with some overlap. Median duration of antifungal exposure was 14 days prior to positive culture. Seven patients (54%) had an open sternum at the time of ECMO cannulation, while 11 (85%) had femoral ECMO access. Four patients (31%) had Candida species isolated from other sources prior to fungemia (mediastinum and respiratory system). Candida parapsilosis (31%), C. albicans (23%), and C. guilliermondii (15%) were the most common isolates. The most likely sources were vascular access catheters (38%) and mediastinum (31%). Seven patients (54%) had suspected or confirmed seeding of ECMO hardware after positive blood cultures, though only two circuit exchanges were performed. Four patients (31%) showed evidence of non-susceptibility to their prophylactic antifungal class (Table 1), with a median 35-day exposure prior to positivity among those patients. Only 62% of patients demonstrated blood culture clearance (median 3 [IQR 2.5-7.75] days). Nearly all patients (92%) died within the hospital stay, including all patients with resistant organisms. Conclusions In an ECMO center with routine antifungal prophylaxis practices, the rate of fungemia identified during ECMO support was low. When fungemia developed, there was a high proportion of drug resistance and mortality. This abstract is funded by: None

Abstract Rationale Longer duration and severity of acute brain dysfunction (ABD), including delirium and coma, have been associated with worse clinical outcomes in critically ill adults. Examining ABD trajectories over time may offer deeper insights into understanding its heterogeneity. This study aimed to characterize ABD trajectories and examine its relationship with clinical characteristics and 30-day mortality. Methods We analyzed data from the multicenter BRAIN-ICU and MIND-ICU prospective cohort study, which included adults with acute respiratory failure and/or shock. Delirium presence and severity were determined twice daily for 7 days using the CAM-ICU-7. A score of 0 indicates no delirium, 1-2 indicates subsyndromal delirium, 3-5 indicates mild-to-moderate delirium, and 6-7 indicates severe delirium. Patients with a RASS of -4 or -5, indicating coma, were assigned a CAM-ICU-7 score of 7. We applied group-based trajectory modeling of CAM-ICU-7 to identify ABD severity trajectory patterns and compared demographics, clinical characteristics, and 30-day mortality among trajectory groups. Results The 1035 patients analyzed had a median age of 62 [IQR, 53-72] years, 61% were male, and 91% were White. Overall, patients had a median Charlson Comorbidity Index of 2 (1-4) and SOFA score at enrollment of 6 (4-9). 731 patients (70.6%) had delirium at some point during the study. Five ABD severity trajectories were identified as the optimal model fit based on global fit indices and clinical interpretability: T1 was associated with mild and persistent ABD, while T2-T5 were characterized by severe ABD early during critical illness with different resolution patterns (Figure). Trajectory membership varied by patient clinical characteristics. T1 had lighter levels of sedation (median RASS score: -2) in the 24 hours surrounding ICU admission than T2-T5 (median RASS scores: -3, -3, -4, and -3, respectively) (p<.001). At enrollment, T1 and T2 had lower modified SOFA scores (median scores: 6) compared to T5 (median score: 7) (p<.001). Trajectory membership independently predicted 30-day mortality, with T5 having the highest 30-day mortality and T2, which had the fastest resolution of ABD, the lowest (p<.001). Conclusions We identified five trajectories of ABD severity over the first week of a patient’s ICU stay. These trajectories were differentially associated with mortality and clinical characteristics, suggesting trajectory-based approaches may enable more nuanced risk stratification and support the tailoring of interventions and clinical trial designs to specific ABD trajectory patterns. Future work should validate these trajectories in independent cohorts, explore whether they modify treatment effects, and determine if they represent distinct mechanistic pathways. This abstract is funded by: NIH/NIA