Citations

Abstract Interstitial lung disease (ILD) is a chronic, irreversible condition that is challenging to diagnose due to subtle symptoms and limitations of conventional tools. Noninvasive lung sound analysis using phonopulmography offers a promising alternative. This study evaluated the feasibility of a deep learning model for distinguishing normal and pathological lung sounds in ILD. Lung sound recordings were collected from 12 subjects (6 healthy, 6 ILD) and segmented beat-to-beat using simultaneously acquired ECG and phonopulmography, yielding 2,280 segments. Mel-frequency cepstral coefficients (MFCCs; 13 per segment) were extracted, and a one-dimensional CNN-based You Only Listen Once (YOLO) model was trained for segment classification. Evaluation used Leave-One-Out Cross-Validation, with patient-level classification by aggregating segment predictions via majority vote. On the test set (190 segments per class), the model achieved an average segment-level accuracy of 0.738 ± 0.087 and mean AUC of 0.850 ± 0.062. Precision, recall, and F1-score were 0.775 ± 0.148, 0.755 ± 0.183, and 0.738 ± 0.089, respectively. These preliminary results suggest that beat-level MFCCs with a 1D CNN-based YOLO model enable accurate segment- and patient-level classification. This feasibility study highlights the potential of AI-assisted auscultation for noninvasive monitoring and clinical decision support in ILD management, pending further validation in larger cohorts.

Abstract Introduction: Asynchronous online courses are plagued with the challenge of engaging students consistently. Behavioral interventions, such as bottlenecks (enforced deadlines) and nudges (automated reminders), may modulate behaviors that lead to engagement and can now be easily applied and measured in a Learning Management System (LMS). Methods: This study included three groups of three asynchronous online courses each over three academic quarters. Two distinct behavioral interventions, bottlenecks and nudges, were implemented to increase engagement with the course compared to non-intervention controls. Results: The results showed that behavioral intervention tools increase student engagement with course materials by measures of increased student access to the course and reduced time to quiz completion. However, there was no significant impact in grades across the three groups. Conclusion: The use of nudges and bottlenecks as behavioral interventions to increase engagement in asynchronous courses for Clinical and Translational Science Education is recommended as part of best practice in course design with the same or similar structure.

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

Key Points In an observational cohort of mostly non-Hispanic White women, pregnancy was not associated with later-life accelerated kidney function decline. Pregnancy occurring during the HALT Progression of Polycystic Kidney Disease (HALT-PKD) trial was not associated with kidney disease progression relative to women who did not become pregnant. Prior evidence indicates that adverse pregnancy outcomes may be involved in autosomal dominant polycystic kidney disease progression; pregnant women should be closely monitored. Background Data are conflicting regarding whether pregnancy influences disease progression in women with autosomal dominant polycystic kidney disease (ADPKD). This study examined whether pregnancy or number of pregnancies was associated with kidney disease progression in women with ADPKD. Methods Women with early- (study A) and late-stage (study B) ADPKD from the HALT Progression of Polycystic Kidney Disease (HALT-PKD) trial were included to examine the association between self-reported number of pregnancies (categorical predictor: no pregnancy versus 1–2 pregnancies and ≥3 pregnancies), annual slope of eGFR, annual percent change in total kidney volume (%ΔTKV), and a composite outcome (kidney failure, 50% decline in eGFR, or death) using multivariable linear regression and Cox proportional hazard models. In addition, women who became pregnant, had full-term pregnancies, and available data during study participation ( n =13) were propensity matched (1:4) to women who were not pregnant during study participation, and a mixed model was applied to determine the association of pregnancy with eGFR slope and %ΔTKV. Results Across all analyses, 455 women with a median age of 45 (interquartile range, 38–50) years and eGFR of 69±25 ml/min per 1.73 m 2 at baseline were included. One hundred ninety-nine women had 1–2 pregnancies, and 165 women had ≥3 pregnancies. There was no association of 1–2 pregnancies or ≥3 pregnancies (versus no pregnancies) with eGFR slope ( β -estimate [95% confidence interval (CI)]; 1–2 pregnancies: 0.22 [−0.44 to 0.89]; ≥3 pregnancies: −0.46 [−1.16 to 0.25]), %ΔTKV ( β -estimate [95% CI]; 1–2 pregnancies: 0.92 [−0.34 to 2.17]; ≥3 pregnancies: 0.69 [−0.67 to 2.04]), or time to composite outcome (hazard ratio [95% CI]; 1–2 pregnancies: 1.04 [0.56 to 1.93]; ≥3 pregnancies: 1.48 [0.78 to 2.77]) in adjusted models. Moreover, there was no difference in annual eGFR slope ( β -estimate: −0.14 [95% CI, −2.72 to 2.44]) and %ΔTKV ( β -estimate: 0.04 [95% CI, −3.74 to 3.82]) in women who became pregnant matched to women who did not became pregnant during their participation in the study. Conclusions Pregnancy was not associated with ADPKD progression among women with early- and late-stage ADPKD enrolled in the HALT-PKD trial.

Background: Vascular endothelial dysfunction is a key antecedent to cardiovascular diseases (CVD) in older adults. Reduced nitric oxide (NO) bioavailability secondary to excess mitochondria-related reactive oxygen species (mitoROS) bioactivity are key mechanisms of endothelial dysfunction with aging. Cellular senescence is a state of largely irreversible cell cycle arrest that drives endothelial dysfunction. This occurs in part through secretion of bioactive factors into serum, collectively referred to as the circulating senescence-associated secretory phenotype (SASP), that promote excess mitoROS and reduce NO. We have shown fisetin improves endothelium-dependent dilation in old mice by reducing cellular senescence and mitoROS. Here, we sought to translate our preclinical findings to older adults. Hypotheses: We tested the primary hypothesis that intermittent fisetin treatment would improve endothelial function in older adults and that this would be mediated by reductions in mitoROS and associated with lower cell senescence. We also determined if fisetin altered the serum to improve endothelial cell (EC) NO and reduce whole-cell ROS and mitoROS ex vivo. Exploratory analyses assessed if fisetin changed circulating SASP factors. Methods: In a randomized, placebo-controlled, parallel-group design pilot trial, 41 older men and postmenopausal women were randomized to fisetin (2 mg/kg/day) (age: 71±1 yrs [mean±SEM], n=22 [13 W]) or placebo (age: 71±1 yrs, n=19 [12 W]) administered in an intermittent manner (2, 3-day dosing periods separated by 14 days). Outcomes were measured at baseline and 1-3 weeks after the final dose. Endothelial function was assessed as brachial artery flow-mediated dilation (FMDBA). The change in FMDBA with an acute supratherapeutic dose (160mg) of the mitochondrial antioxidant MitoQ was used to assess mitoROS-related inhibition of endothelial function. Gene expression of cellular senescence markers Cdkn1a (encodes p21) and Lmnb1 (encodes Lamin B1) were assessed in peripheral blood mononuclear cells (PBMCs) by Q-PCR. Human aortic ECs (HAECs) in culture were exposed to 10% participant serum for 2 hrs; acetylcholine (ACh)-stimulated NO production (DAR-4M-AM) and basal whole-cell ROS (CellROX) and mitoROS (MitoBright Deep Red) bioactivity were assessed via immunofluorescence. Luminex analyses (Biotechne) measured circulating levels of SASP factors in serum. Results: FMDBA was improved by 41% with fisetin (pre: 3.9±0.4%, post: 5.5±0.5%, P< 0.0001), but unchanged with placebo (pre: 4.4±0.5%, post: 4.5±0.4%) (group x time interaction effect: P< 0.01). Similar efficacy was observed in men and women. Acute MitoQ enhanced FMDBA at baseline and after placebo (all P< 0.05) but led to no further improvements in FMDBA after fisetin treatment. Consistent with lower cell senescence, expression of Cdkn1a was reduced (P< 0.05) by 12% and expression of Lmnb1 increased (P=0.06) by 16% in PBMCs collected after fisetin but not placebo. Compared with pre-intervention, serum collected after fisetin, but not placebo, increased HAEC ACh-stimulated NO by 6% (P< 0.01) and lowered whole-cell and mitoROS bioactivity by ~10% (both P< 0.05). Protein levels of CXCL12 in serum were 5% lower (P< 0.05) after fisetin compared with pre-intervention and after placebo. Traditional cardiometabolic risk factors were unchanged. Conclusions: Intermittent fisetin treatment improves vascular endothelial function in older adults in part by suppressing mitoROS. These effects are linked to reduced cell senescence and select changes to the circulating SASP. Funding: AHA 23CDA1056582; NIH F31AG087709; NIH/NCATS CTSA UL1TR002535; Life Extension® provided fisetin (BioFisetin) and placebo. This abstract was presented at the American Physiology Summit 2026 and is only available in HTML format. There is no downloadable file or PDF version. The Physiology editorial board was not involved in the peer review process.