Browse Articles
Discover research articles across all indexed journals
Scaling language model size yields diminishing returns for single-message political persuasion
Large language models can now generate political messages as persuasive as those written by humans, raising concerns about how far this persuasiveness may continue to increase with model size. Here, we generate 720 persuasive messages on 10 US political issues from 24 language models spanning several orders of magnitude in size. We then deploy these messages in a large-scale randomized survey experiment ( N = 25,982) to estimate the persuasive capability of each model. Our findings are twofold. First, we find evidence that model persuasiveness is characterized by sharply diminishing returns, such that current frontier models are only slightly more persuasive than models smaller in size by an order of magnitude or more. Second, we find that the association between language model size and persuasiveness shrinks toward zero and is no longer statistically significant once we adjust for mere task completion (coherence, staying on topic), a pattern that highlights task completion as a potential mediator of larger models’ persuasive advantage. Given that current frontier models are already at ceiling on this task completion metric in our setting, taken together, our results suggest that further scaling model size may not much increase the persuasiveness of static LLM-generated political messages.
Abstract 053: Menopausal Status and Race/Ethnicity on the Association of Lipoprotein(a) with CAD
Overview: Lipoprotein(a) (Lp(a)) is a risk factor for atherosclerotic coronary artery disease (CAD). Little is known about menopausal status and hormone therapy in racial/ethnic-specific associations of elevated Lp(a) with CAD. We investigated this in an age- and racially-diverse cohort of women from the U.S.’s largest integrated healthcare system. Methods: We extracted structured cross-sectional electronic health record data for 2831 women with a laboratory result for Lp(a) between 1999-2023 in the Veterans Health Administration. We designated menopausal status (pre, surgical, natural), use of hormone contraception or menopausal hormone therapy (HT), and history of CAD as of the Lp(a) test date. Elevated Lp(a) was Lp(a) >125 nmol/L. Using multivariate logistic regression with correction for multiple comparisons, we estimated association of elevated Lp(a) with prevalent CAD adjusted for age, race/ethnicity, menopausal status, and HT. Within menopausal subgroups, we conducted analogous multivariate logistic regressions. We tested for interaction between Lp(a) and menopausal status and between HT and race/ethnicity. Results: Elevated Lp(a) was associated with prevalent CAD among all women (OR=1.5, 95% CI [1.2, 1.9], p<0.002). In subgroup analysis by menopause status, elevated Lp(a) was associated with prevalent CAD only for women over 60 with natural menopause (OR=2.1, 95% CI [1.5, 3.0], p<0.001). An interaction model between elevated Lp(a) and menopause status on prevalent CAD was assessed with premenopausal women under 60 without elevated Lp(a) as the reference group. Premenopausal women under 60 with elevated Lp(a) had an OR of 1.4 (95% CI [0.8, 2.5] p=0.274). Women over 60 with natural menopause and without elevated Lp(a) had an OR of 4.9 (95% CI [3.4, 7.3], p<0.001). Women over 60 with natural menopause and elevated Lp(a) had the highest OR of 10.4 (95% CI [6.7, 16.0], p<0.001). Relative excess risk due to interaction (RERI) of 5.1 (95% CI [2.2, 9.7]) supports an additive interaction of menopausal status over/under age 60 with Lp(a) on risk of prevalent CAD. The p-value for multiplicative interaction was 0.22. In smaller-sized menopausal subgroups, no observations were observed between race/ethnicity and HT on risk of elevated Lp(a) or on prevalent CAD. Conclusions: Menopausal status over age 60 and elevated Lp(a) (>125 nmol/L) interacted to modify the risk of prevalent CAD. HT use as a possible effect modifier, by race/ethnicity, warrants further exploration.
Abstract MP01: Butter and Plant-Based Oils Intakes and Mortality in Three Large Prospective Cohorts of US Women and Men
Objective: To investigate associations of butter and plant-based oil intakes with risk of total and cause-specific mortality. Participants: 221,054 women and men from the Nurses’ Health Study (1990-2018), Nurses’ Health Study II (1991-2018), and Health Professionals Follow-up Study (1990-2018) were included, all of whom were free of cancer, cardiovascular disease (CVD), diabetes, or neurodegenerative disease at baseline. Exposure: We assessed dietary intake repeatedly using validated semiquantitative food frequency questionnaires every 4 years. Primary exposures included total butter (butter added at the table and from baking and frying) and plant-based oil intake (safflower, soybean, corn, canola, and olive oil). Main Outcome: We identified deaths through the National Death Index and other sources. A physician classified the cause of death based on all available records. Total mortality was the primary outcome, and mortality due to cancer and CVD were secondary outcomes. Results: During up to 28 years of follow-up, we documented 41,618 deaths, including 11,380 from cancer and 9,114 from CVD. After adjusting for confounding factors, the highest butter intake was associated with an 18% higher risk of total mortality compared to the lowest intake (HR: 1.18; 95% CI: 1.10-1.25; P trend <0.001). Conversely, the highest plant-based oil intake was linked to an 18% lower risk of total mortality (HR: 0.82; 95% CI: 0.76-0.89; P trend <0.001). Higher olive, soybean, and canola oil intakes were also significantly associated with lower total mortality (all P trend <0.001). For every 10 g/day increase in plant-based oil intake, cancer mortality risk decreased by 11% (HR: 0.89; 95% CI: 0.85-0.94; P trend <0.001), and CVD mortality risk by 6% (HR: 0.94; 95% CI: 0.89-0.99; P trend = 0.03). Higher butter intake was associated with increased cancer mortality (HRper10g/day: 1.12; 95% CI: 1.04-1.20; P trend <0.001), but not with CVD mortality ( P trend =0.61). Replacing 10 g/day of butter with plant-based oils was associated with a 19% reduction in total mortality (HR: 0.81; 95% CI: 0.77-0.85; P trend <0.001) and a 20% reduction in cancer mortality (HR: 0.80; 95% CI: 0.74-0.88; P trend <0.001). Conclusion: A higher intake of butter is associated with higher mortality, while a higher intake of plant-based oil is associated with lower mortality. Substituting butter with plant-based oils may confer substantial benefits for preventing premature deaths.
Abstract P3120: Obesity and the Risk of Advanced Stages of Cardiovascular-Kidney-Metabolic Syndrome among Obstructive Sleep Apnea Patients
Background: The cardiorenal metabolic (CKM) syndrome is a condition characterized by the interrelationship between the risk factors and relevant comorbidities. Obstructive sleep apnea (OSA) is associated with an increased risk of the advanced stages of CKM, cardiovascular disease (CVD) and chronic kidney disease (CKD). Obesity has been recognized as a significant risk factor for both the onset and worsening of OSA. However, it remains unclear whether the severity of obesity correlates with varying degrees of risk for CVD and CKD among patients with OSA. Methods: Using the most recent dataset from the All of Us Research Program, we included 29,570 participants with a BMI of ≥18.5 kg/m 2 , accessible electronic health records (EHR), and no prior history of OSA. BMI was assessed during physical examinations, and participants were categorized into normal weight, overweight, Class I, Class II, and Class III obesity based on World Health Organization criteria. We utilized the Systematized Nomenclature of Medicine to ascertain outcomes from the EHR, with CVD defined as a composite of heart failure, myocardial infarction, stroke, and atrial fibrillation. Cox regression models, adjusted for age, sex, race, and education, were employed to calculate hazard ratios (HRs) for CVD and CKD across the BMI categories in individuals with OSA. Results: The mean age of participants was 59.8 years (SD: 13.1), with females comprising 51.9% of the total sample. 70.7% were classified as with obesity. In general, compared to those with normal weight, participants with classes of obesity exhibited an increased risk of CVD and CKD; however, no association was observed between class I obesity and CVD. The risk increased with obesity severity. Class III obesity posed the highest risk, with a 2.34-fold increased risk of CVD (HR: 2.34, 95% CI: 1.87–2.93), a 2.38-fold increased risk of CKD (HR: 2.38, 95% CI: 1.82–3.10). Participants with overweight showed no statistically significant associations with incident CVD or CKD compared to those of normal weight. Conclusions: In patients with OSA, the risk of advanced stages of cardiorenal metabolic (CKM) syndrome, specifically CVD and CKD, progressively increases with more severe obesity. Our findings highlight the focused and effective weight management strategies, in addition to the regular treatment of OSA, to mitigate long-term CKM complications, ultimately aiding in the reduction of the complex and detrimental interplay of CKM.
Visualizing agonist-induced M2 receptor activation regulated by aromatic ring dynamics
Despite the growing number of G protein–coupled receptor (GPCR) structures being resolved, the dynamic process of how GPCRs transit from the inactive toward the active state remains unclear. In this study, comprehensive molecular dynamics simulations were performed to explore how ligand binding modulates the conformational dynamics of the M2 muscarinic acetylcholine receptor (M2R). We observed a sequential occurrence of structural changes in the inactive-to-active transition of M2R induced by a superagonist iperoxo, which includes the orthosteric binding site contraction, the TM6 opening into an intermediate conformation, and a further structural change toward full activation upon binding to G protein or a G protein mimetic nanobody. Two activation intermediates were identified, which show structural features different from those reported for apo-GPCRs. Moreover, our results suggest that stabilization of a specific W400 6.48 conformation and enhanced F396 6.44 dynamics are crucial for activation, whereas distinct side-chain rotamer equilibriums of Y206 5.58 in the cytoplasmic cavity are correlated with agonist efficacies. Our work provides atomic-level structural insights into the agonist-induced M2R activation pathway and highlights a mechanism by which ligand efficacy can be encoded and transduced in the form of aromatic ring dynamics.
Abstract 068: Associations Between Fish Oil Supplement Use or Plasma Omega-3 Levels with Risk for Atrial Fibrillation: The United Kingdom Biobank
Background: Recent observational studies in the UK Biobank (UKBB) concluded that self-reported fish oil supplement (FOS) use is associated with an increased risk for incident atrial fibrillation (AF). This lies in contradiction with a globally representative meta-analysis, which found an inverse relationship between blood levels of omega-3 fatty acids (n-3 FAs) and risk of AF. The extent to which plasma levels of n-3 FAs are related to risk of AF in UKBB has yet to be reported. Objectives: We have leveraged data from the UKBB to 1) determine the relationship between plasma levels of n-3 FAs and incident AF and 2) to further explore the previously reported association between FOS use and incident AF. Methods: Within the UKBB, we identified 266,477 individuals with data on blood plasma n-3 FAs and relevant covariates, and 433,607 individuals with data on self-reported FOS use. The primary outcome was incident AF during the follow-up period (median 12.7y). Multivariable-adjusted hazard ratios (95% CIs) for FAs were computed continuously (per inter-quintile range [IQ 5 R]) and by quintile (Q). HRs were computed for dichotomous FOS use. Covariates included: age, sex, ethnicity, education, physical exercise, smoking, alcohol use, BMI, use of beta-blocker, drugs for hypertension or cholesterol, prevalent diabetes, CVD or heart failure, and plasma linoleic acid levels. Notably, in our analyses we adjusted for age as a continuous variable to more completely account for age-related risk of AF, compared to previous analyses which adjusted for age as a dichotomous variable (i.e., 65+ vs <65) in their assessment of FOS and risk of AF. Results: Total n-3 levels in blood plasma were inversely associated with incident AF (HR per IQ 5 R = 0.90 [95% CI 0.86, 0.93]), and HR=0.87 (0.83, 0.91) in Q5 (vs Q1). FOS use was reported by 31% of the cohort, with higher use reported in older individuals. After adjusting for age continuously, there was no association between FOS use and risk of AF risk (HR=1.00 [097, 1.02]). Conclusion: In agreement with recent biomarker-based meta-analyses, higher circulating blood levels of n-3 FA were associated with reduced risk for AF in the UKBB. Secondly, this study reassessed the relationship between FOS use and risk of AF in the UKBB, and if age is adjusted for in a continuous fashion, the association between FOS and AF disappears. These findings indicate previous analyses may have insufficiently adjusted for the age-related risk of AF.
Abstract MP22: A Measurement Model of Socioeconomic Status and its Association with Cardiovascular Disease in the Hispanic Community Health Study/Study of Latinos
Objectives: Low socioeconomic status (SES) is consistently associated with adverse cardiovascular health. Three key indicators are typically used as proxies for SES: income, education, and occupation. However, these indicators may not fully capture the unique sociodemographic features relevant to the SES of Hispanics/Latinos in the U.S. This study aims to expand the traditional SES model by identifying and incorporating additional features specific to Hispanic/Latino persons, providing a more comprehensive assessment of SES and its relationship with cardiovascular disease (CVD). Methods: The Hispanic Community Health Study/Study of Latinos (HCHS/SOL) is a community-based longitudinal cohort study of 16,415 adults self-identifying as Hispanic/Latino, enrolled from 2008-2011 from four U.S. urban communities. We utilized a multiple indicator multiple cause (MIMIC) model to identify formative and reflective indicators of a latent variable for SES. CVD was ascertained by self-report during the second in-person clinic examination from 2014-2017 (V2; N=11,623) and was defined as having had a heart attack, stroke, or angioplasty, stent, or bypass procedure. A survey logistic regression model was then used to examine the association of the latent SES with CVD at V2. Models were adjusted for age, sex, and study site. Results: Significant formative indicators of the SES latent variable included education level, whether the highest level of education was obtained in the US, employment status, and age at immigration (included in a second model with immigrants only; n=9,623). Significant reflective indicators of the SES latent variable included income, the MacArthur SES ladder, and affluence level. The measurement model demonstrated good fit to the data (RMSEA=0.016; CFI=0.984, SRMR=0.02). The SES latent variable was associated with CVD (OR=0.70, 95% CI 0.58, 0.84; for immigrants only OR=0.64, 95% CI 0.53-0.77). Conclusions: Incorporating indicators relevant to Hispanic/Latino populations provides a more comprehensive assessment of SES with potentially improved validity. The latent SES variable, including factors such as education location and age at immigration, was significantly associated with lower odds of cardiovascular disease. These findings highlight the importance of using a culturally-tailored SES framework when examining health outcomes in Hispanics/Latinos.
Abstract P3157: Traumatic Brain Injury and Risk of Early Onset Dementia: A Population-Based Cohort Study and Meta-analysis.
Introduction: Traumatic brain injury (TBI) is a recognized risk factor for late-onset dementia (LOD), though its specific association with early-onset dementia (EOD) (i.e., diagnosed before age 65) is less understood. This study examines the association between TBI, and the risk of developing EOD compared to LOD through a population-based cohort study. Additionally, a systematic review with meta-analysis examines the association between TBI and EOD. Hypothesis: TBI is a risk factor for developing EOD and this association between TBI and EOD is stronger than the association with LOD. Methods: We utilized data from 502,039 participants aged 40 - 69 years at recruitment from the UK Biobank and found that 16,959 people had TBI. Over a median follow-up of 13.6 years, 837 people were diagnosed with EOD, and 8948 with LOD. We estimated hazard ratios (HRs) for TBI's association with EOD and LOD using Cox proportional hazard models with TBI modeled as both a time-varying exposure and time-varying coefficient. Additionally, we conducted a systematic review of existing studies of TBI and EOD, with pooled effect estimates generated via a random-effects model. Results: TBI was significantly associated with an increased risk of both EOD (HR: 3.3, 95% CI: 2.5-4.5) and LOD (HR: 2.5, 95% CI: 2.3-2.7). The association was stronger for EOD compared to LOD (p-value: 0.05), especially in the case of moderate to severe TBI (p-value: 0.15) (Figure 1). A meta-analysis including 10 studies and the present study using UK Biobank, estimated a pooled effect estimate of 1.9 (95% CI: 1.2-3.0), with effect estimates ranging from 0.7 to 5.4, for EOD associated with TBI, albeit with high heterogeneity (I 2 = 96%). Conclusion: This study underscores that TBI is associated with an increased risk of EOD, more so than LOD. These findings emphasize the need for targeted preventive strategies and interventions for individuals with a history of TBI to mitigate or delay dementia onset.
Bacterial estrogenesis without oxygen: Wood–Ljungdahl pathway likely contributed to the emergence of estrogens in the biosphere
Androgen and estrogen, key sex hormones, were long thought to be exclusively produced by vertebrates. The O 2 -dependent aromatase that converts androgen to estrogen (estrogenesis) has never been identified in any prokaryotes. Here, we report the finding of anaerobic estrogenesis in a Peptococcaceae bacterium ( Phosphitispora sp. strain TUW77) isolated from the gut of the great blue-spotted mudskipper ( Boleophthalmus pectinirostris ). This strain exhibits testosterone fermentation pathways, transforming testosterone into estrogens and androstanediol under anaerobic conditions. Physiological experiments revealed that strain TUW77 grows exclusively on testosterone, utilizing the androgenic C-19 methyl group as both the carbon source and electron donor. The genomic analysis identified three copies of a polycistronic gene cluster, abeABC (anaerobic bacterial estrogenesis), encoding components of a classic cobalamin-dependent methyltransferase system. These genes, highly expressed under testosterone-fed conditions, show up to 57% protein identity to the characterized EmtAB from denitrifying Denitratisoma spp., known for methylating estrogen into androgen (the reverse reaction). Tiered transcriptomic and proteomic analyses suggest that the removed C-19 methyl group is completely oxidized to CO 2 via the oxidative Wood–Ljungdahl pathway (WLP), while the reducing equivalents (NADH) fully reduce remaining testosterone to androstanediol. Consistently, the addition of anthraquinone-2,6-disulfonate, an extracellular electron acceptor, to testosterone-fed TUW77 cultures enabled complete testosterone conversion into estrogen without androstanediol accumulation (anaerobic testosterone oxidation). This finding of aromatase-independent estrogenesis in anaerobic bacteria suggests that the ancient WLP may have contributed to the emergence of estrogens in the early biosphere.
Abstract P2167: Racial and Ethnic Disparities in Neighborhood Socioeconomic Status Associated with Cardiovascular Disease and Stroke
Introduction: The role of socioeconomic status (SES) across cardiovascular diseases (CVD) is well researched at the individual-level but fails to explain the greater CVD rates among Black and American Indian and Alaska Native (AIAN) groups at every level of SES. This research investigated whether lower Neighborhood SES (NSES), a structural determinant of health, was associated with racial and ethnic disparities in CVD or stroke among aging women. Methods: The data came from the Women’s Health Initiative (WHI) cohort, consisting of 161,808 postmenopausal women with 30 years of data on stroke and CVD incidence. The NSES data were obtained from the US Census based on participant’s census tract of residence. We evaluated racial and ethnic disparities in the exposure and outcome prevalence using ANOVA. Associations between NSES and time to stroke and CVD were evaluated using multivariate adjusted Cox Regression models, and models were stratified by race and ethnicity to evaluate disparities. Results: NSES differed by race and ethnicity. Black, AIAN, and Latina groups were significantly (P<0.01) more likely to live in neighborhoods with NSES Z-scores below the population mean (Z-score=0) compared with White and Asian groups. The mean NSES Z-score among Black participants was -5.21 standard deviations below the population mean. Thirty-year CVD and stroke incidence followed the same trajectory of racial disparities. Preliminary analysis suggests results from Cox regression models will identify lower NSES to be longitudinally associated with greater stroke and CVD incidence overall, and greater effect sizes among groups experiencing structural racism. Conclusion: This research evaluates racial disparities in NSES as a structural determinant of stroke and CVD. These findings provide evidence emphasizing the urgent need for neighborhood-level interventions to address CVD disparities that coalesce in marginalized communities.
Abstract P1065: Contemporary trends and in-hospital outcomes of infective endocarditis among individuals with substance use disorders: A nationwide analysis
Background: Infective endocarditis (IE) is a serious complication linked to substance use disorders (SUD). This study aimed to examine the contemporary nationwide trends and outcomes of IE hospitalizations among individuals with SUD. Methods: We identified hospitalizations with any IE diagnosis from the Vizient ® Clinical Data Base between 2016-2023, data used with permission of Vizient, Inc. (All rights reserved). We examined IE trends in SUD vs. non-SUD individuals and used Cochrane-Armitage Test to analyze significance. Multivariable logistic regression was used to assess variables associated with in-hospital mortality among individuals with IE. Results: Among 159,045 individuals hospitalized with IE, 37,596 (23.6%) had a history of SUD. Individuals with IE associated SUD vs. non-SUD were younger (39 vs. 63 years), predominantly White (77.3%), and primarily on Medicaid (66.5%). Staphylococcus species caused 53.2% of IE associated SUD hospitalizations. IE-SUD hospitalizations rose from 58.8 per 100,000 in 2016 vs 70.7 per 100,000 in 2022, then declined to 56.7 in 2023, P<0.001. Non-SUD hospitalizations for IE increased from 233.9 per 100,000 in 2016 vs 262.9 per 100,000 in 2023, P<0.001(Figure). In-hospital mortality was lower in the IE-SUD group vs. IE non-SUD group (8.6% vs 12.8%, P<0.001). The following factors were independently associated with in-hospital mortality: older age (OR 1.016, 95% CI 1.015–1.018), fungemia (OR 2.43, 95% CI 2.20–2.68), and hemodialysis (OR 3.44, 95% CI 3.29–3.59). Conclusion: This contemporary nationwide observational analysis highlights the evolving national trends and outcomes of IE associated with SUD. Future research should explore the drivers of these trends to guide policy, with a focus on enhanced screening and preventative measures.
Abstract P3155: Socio-Demographic Disparities in the Management Modalities of Pulmonary Embolism Amongst COVID-19 Patients: A Retrospective Population-Based Study
Introduction: Pulmonary embolism (PE) affects 60 to 70 persons per 100,000 people and causes about 7 million DALYs worldwide, making it an economically onerous handicap condition in the US. COVID-19's hypercoagulation and pro-inflammatory condition increase PE risk (OR 4.4 and RR 3.1) compared to non-COVID-19 individuals. Few studies have examined modalities and disparities in use. Aims: The study's main goal was to calculate the utilization rate and predictors of utilization of PE therapy modalities (tPA/thrombolytics, mechanical thrombectomy [MT], and surgical thrombectomy [ST]), and treatment disparities in COVID-19 patients in the US. Methods: We performed a retrospective cross-sectional observational study of a nationwide inpatient sample (year 2020) in adult hospitalizations with PE to identify patients with COVID-19. ICD-10 codes were used. Using SAS 9.4, we ran a univariate analysis using the chi-square test and multivariate survey logistic regression analysis to calculate adjusted OR and 95% CI. Results: Of 172,630 PE hospitalizations, 3124 patients had COVID-19. PE patients with COVID-19 were younger (mean age 60 vs 66) with the age group of 18-44 (2.54% vs >65-year-old 1.36%), male (1.99% vs female 1.63%), and Hispanics (4.47% vs African Americans 2.71% vs White 1.35%) in comparison to non-COVID-19. PE with COVID-19 had a higher prevalence of obesity (30.72% vs 29.6%) and diabetic mallitus (12.8% vs 9.76%). Utilization of tPA (3.36% vs 3.14%), MT (7.04% vs 5.58%), and ST (0.32% vs 0.19%) were higher amongst PE with COVID-19. tPA use was higher amongst young [OR 1.47 (95%CI 1.01-2.14), females (1.54, 1.17-2.04), and patients with diabetes mellitus (1.73, 1.18-2.53). MT use was higher amongst young (1.46, 1-2.12), African Americans (1.8, 1.45-2.24) and Hispanics (2.99, 2.06-4.35) (compared to White), patients with cardiogenic shock (3.06, 1.91-4.90), ischemic stroke (3.33, 1.55-7.15), diabetes mellitus (1.29, 1.01-1.63), and obesity (2.54, 2.15-3.0). ST use was higher amongst COVID-19 (7.48, 1.94-28.88), males (17.54, 3.5-83.33), patients with AFib (7.49, 1.68-33.49), and obesity (4.75, 1.37-16.51). Conclusion: Management discrepancies among PE patients with COVID-19 include younger age, male, African American, and Hispanic, with ischemic stroke, AFib, obesity, and diabetes using more. To reduce socio-demographic differences in COVID-19 pulmonary embolism care, personalized management techniques, and equitable healthcare resources should be developed.
Candidate transmission survival genome of <i>Mycobacterium tuberculosis</i>
Mycobacterium tuberculosis (Mtb), a leading cause of death from infection, completes its life cycle entirely in humans except for transmission through the air. To begin to understand how Mtb survives aerosolization, we mimicked liquid and atmospheric conditions experienced by Mtb before and after exhalation using a model aerosol fluid (MAF) based on the water-soluble, lipidic, and cellular constituents of necrotic tuberculosis lesions. MAF induced drug tolerance in Mtb, remodeled its transcriptome, and protected Mtb from dying in microdroplets desiccating in air. Yet survival was not passive: Mtb appeared to rely on hundreds of genes to survive conditions associated with transmission. Essential genes subserving proteostasis offered most protection. A large number of conventionally nonessential genes appeared to contribute as well, including genes encoding proteins that resemble antidesiccants. The candidate transmission survival genome of Mtb may offer opportunities to reduce transmission of tuberculosis.
Abstract P1110: Pecan Intake Improves Lipoprotein Particle Concentrations Compared to Usual Intake in Adults at Increased Risk for Cardiometabolic Diseases: a 12-Week Randomized Controlled Trial
Background: Pecan consumption consistently improves lipids and lipoproteins, but less research has investigated the effect of pecan consumption on lipoprotein subfractions. Objectives: The aim was to investigate how substitution of usual snack foods with 57 g/day of pecans affects the concentration of lipoprotein particle subfractions, lipoprotein particle size and apolipoproteins compared to continuing usual intake after 12 weeks in adults at risk for cardiometabolic diseases. Exploratory analyses were done to evaluate early markers of insulin resistance including the Lipoprotein Insulin Resistance score (LPIR), Diabetes Risk Index (DRI), and GlycA. Methods: This is a secondary analysis of data from a 12-week randomized controlled trial including adults at risk for cardiometabolic disease. Participants were randomized to either consume 57 g/day of pecans in place of usual snacks or to continue their usual intake. EDTA plasma samples collected at baseline and after the 12 week-intervention period were used for analysis. Samples were analyzed via proton nuclear magnetic resonance spectroscopy for lipoprotein subfractions, apolipoproteins, GlycA, and branched chain amino acids. Lipoprotein subfractions and branched chain amino acids were used to calculate the LPIR and DRI. Results: The pecan group had a greater reduction from baseline in the concentrations of apolipoprotein B (apoB) (-4.38 mg/dL; 95% CI -8.02, -0.73), total LDL (-75.3 nmol/L; 95% CI -144, -6.93), total triglyceride rich lipoproteins (TRLP) (-20.4 nmol/L; 95% CI -33.8, -7.03), large TRLP (-1.47 nmol/L; 95% CI -2.69, -0.26), small TRLP (-11.3 nmol/L; 95% CI -22.4, -0.27), and the LPIR (-4.42 points; 95% CI -8.14, -0.69) and greater increases from baseline in the concentration of large high density lipoproteins (HDL) (0.35 μmol/L; 95% CI 0.07, 0.63) and HDL subspecies, H5 (0.40 μmol/L; 95% CI 0.10, 0.62) compared to the usual diet group. There were no between-group differences in branched chain amino acids or the DRI. Conclusions: Incorporating 57 g/day of pecans into the diet in place of usual snacks for 12 weeks reduced apoB, atherogenic lipoprotein subfractions, and an early marker of insulin resistance compared to usual intake.
Abstract P1037: Promoting Sustained Behavior Change and Nutrition Security in Medicaid-Enrolled Individuals with Stage 2 Cardiovascular Kidney Metabolic Syndrome (CKMS): protocol and early recruitment success of the SUSTAIN study
Introduction: Medicaid-enrolled populations are disproportionately impacted by suboptimal nutrition leading to lower Life’s Essential 8 (LE8) cardiovascular health (CVH) scores and greater prevalence of Stage 2 cardiovascular-kidney-metabolic syndrome (CKMS) as defined by the American Heart Association (AHA). With support from the AHA Health Care by Food TM Initiative and leveraging existing academic-community-government-industry partnerships, our team is assessing the feasibility, engagement, and preliminary efficacy of SUSTAIN , a novel Food is Medicine (FIM) intervention in Medicaid-enrolled individuals with Stage 2 CKMS. Hypothesis: Our overall hypothesis is that a comprehensive FIM intervention including culturally appropriate and home-delivered groceries, intensive and tailored behavioral nutrition counseling, and social care coordination will improve nutrition security and CVH compared to individuals receiving home-delivered groceries only. Methods: Informed by the socioecological model and human-centered design, SUSTAIN is a 24-week randomized controlled trial (RCT) of Medicaid-enrolled individuals (n=100) with Stage 2 CKMS to: Aim 1 : Determine the feasibility and engagement of participants in the SUSTAIN intervention compared to enhanced usual care over 24 weeks through mixed-methods measurement of participant enrollment, adherence, retention, and engagement (counseling, screenings, referrals, and uptake); Aim 2 : Determine the preliminary efficacy of the SUSTAIN intervention compared to enhanced usual care in improving behavior change as measured by nutrition security, LE8 measures, biometrics, and purchasing behaviors over 24 weeks to determine effect estimates to robustly power a future RCT. Results: Initial recruitment efforts have enrolled 46 individuals; 148 failed to meet eligibility criteria. Primary reasons for ineligibility were other diagnoses (e.g., heart failure) (n=107) and non-English speaking (n=31). A minority of participants were ineligible due to unwillingness to use mobile phone-based technology (n=2) or lack of backup form of payment for grocery vendor (n=3). Conclusions: Innovative FIM solutions are urgently needed. SUSTAIN will explicate the feasibility, engagement, and preliminary efficacy of SUSTAIN, leading to a scalable FIM nutrition-security-based model to improve CVH in Stage 2 CKMS. SUSTAIN has the potential to drive a paradigm shift in healthcare, fostering healthier, nutrition-secure and resilient communities.
Abstract P1153: Hypoxia Response Markers are Associated with Early Metabolic Imbalance: The U.S. National Health and Nutrition Examination Survey
Introduction: The metabolism spectrum considers cardiometabolic health&disease as a series of four distinct categories, with progressing onset and severity of the underlying pathophysiology (Fig. 1). While prediabetes and metabolic syndrome are widely recognized categories, early metabolic imbalance (EMI) has largely been overlooked. Prevalent in teens and young adults, EMI is characterized by compensated insulin resistance, where fasting glucose, triglycerides, HDL cholesterol (HDL), and hemoglobin A1c are all within normal limits. Thus, EMI does not meet the criteria for prediabetes or metabolic syndrome and eludes risk screening for type 2 diabetes and ASCVD. In EMI, high circulating insulin overproduces nitric oxide, possibly leading to oxidized hemoglobin and a subclinical impairment of oxygen delivery to cells and tissues. Here, we investigated the association of hypoxia response markers with EMI in the U.S. population. Hypothesis: Individuals with EMI show a compensatory response suggestive of subclinical hypoxia. Methods: The fasting subsample from the 2015-2018 U.S. National Health&Nutrition Examination Survey was analyzed: 6,227 observations representing 270.2 million people in the U.S. ages 12 and up. Population-weighted multinomial logistic regression was performed using Stata v18.5. The outcome variable was cardiometabolic health, categorized into 1 of 4 groups: healthy balanced metabolism, EMI, prediabetes/dyslipidemia (PD), or type 2 diabetes/ASCVD (Fig. 1). EMI was defined using prognostic insulin cutpoints derived from timeROC analysis of the CARDIA cohort. The 1° exposure was blood hemoglobin concentration. Resting pulse rate, serum globulin/albumin ratio (chronic inflammation), and serum GGT (cell damage) were 2° exposures. The adjusted covariates are listed in Table 1. To quantify association, the effect size was relative risk ratio with 95% confidence interval and p-value. Results: After adjusting for covariates, markers of the hypoxia response (hemoglobin, resting pulse rate, globulin/albumin ratio, GGT) were elevated in EMI compared with healthy balanced metabolism (Table 1, bottom panel). These increases were not apparent from a simple inspection of the unadjusted median values (top panel). Conclusion: EMI includes an insidious compensatory response to hypoxia. This observation is consistent with prior observations that dysfunctional hemoglobin is coupled to poor metabolic health. Further investigation is warranted.
Structure of activity in multiregion recurrent neural networks
Neural circuits comprise multiple interconnected regions, each with complex dynamics. The interplay between local and global activity is thought to underlie computational flexibility, yet the structure of multiregion neural activity and its origins in synaptic connectivity remain poorly understood. We investigate recurrent neural networks with multiple regions, each containing neurons with random and structured connections. Inspired by experimental evidence of communication subspaces, we use low-rank connectivity between regions to enable selective activity routing. These networks exhibit high-dimensional fluctuations within regions and low-dimensional signal transmission between them. Using dynamical mean-field theory, with cross-region currents as order parameters, we show that regions act as both generators and transmitters of activity—roles that are often in tension. Taming within-region activity can be crucial for effective signal routing. Unlike previous models that suppressed neural activity to control signal flow, our model achieves routing by exciting different high-dimensional activity patterns through connectivity structure and nonlinear dynamics. Our analysis of this disordered system offers insights into multiregion neural data and trained neural networks.
Abstract 062: Disease Trajectories of Cardiometabolic Diseases and Depression: Transition Patterns, Multiomics Signatures, Prognosis and Prediction
Objective: Cardiometabolic diseases (CMDs) and depression, among the most prevalent physical and mental diseases, frequently co-occur and are associated with a higher risk of premature mortality. However, the trajectories of their multimorbidity occurrence, underlying biological mechanisms, and early prediction remain poorly understood. Methods: This study was conducted on 467,592 UK Biobank participants without baseline CMDs and depression. CMDs included type 2 diabetes, coronary artery disease, stroke, and heart failure. Multistate models were used to investigate transition probabilities and identify multiomics signatures for transitions from baseline to single morbidity and to multimorbidity. Multiomics prediction models were constructed using least absolute shrinkage and selection operator Cox regression. Model performance was evaluated by the change of Harrell's C statistic (Δ C-statistic) compared to the demographic model and area under receiver-operating characteristic curves (AUCs). Results: During a median follow-up of 14.6 years, 64,442 participants developed CMDs alone, 17,533 developed depression alone, and 6104 developed multimorbidity. Depression preceding CMDs showed a 2.5% higher 15-year multimorbidity probability and 1.2% higher 5-year mortality risk than the reverse sequence. Multimorbidity was associated with a 14-26% higher mortality risk and 3.6-3.8 years shorter survival time compared to those without CMDs or depression. Distinct and shared multiomics signatures underlying disease trajectories were identified. Valine, leucine, and isoleucine biosynthesis and cytokine-cytokine receptor interaction pathways were implicated in both CMDs and depression progression. Proteomics scores showed superior prediction performance across nine disease transitions, with Δ C-statistic ranging from 0.07 for health-death to 0.22 for health-depression-CMDs-death, compared to the genomics (0-0.04) and metabolomics (0-0.16) scores. For 15-year outcome prediction, the proteomics scores model achieved AUCs of 0.65-0.87 for various transitions, significantly outperforming traditional risk factors-based and other omics models. Conclusions This study revealed distinct transition patterns, identified associated multiomics signatures, and constructed multiomics prediction models for the multimorbidity cluster of CMDs and depression.
Abstract 031: Associations of Adipokines with 6-Year Progression of Cardiovascular-Kidney-Metabolic Syndrome: The Atherosclerosis Risk in Communities (ARIC) Study
Background: Adipose tissue and its signaling molecules are central to cardiovascular-kidney-metabolic (CKM) syndrome development. High levels of pro-inflammatory adipokines and low levels of anti-inflammatory adipokines have been linked with worsening metabolic health and increased cardiovascular disease (CVD) risk. However, there have not yet been investigations of how adipokines relate to changes in CKM health over time. Methods: We examined 7695 White and Black ARIC participants without baseline CVD and with adipokine measurements and data required to assess CKM stage at both Visits 2 (1990-92) and 4 (1996-98), an approximate 6-year interval. CKM stage was defined as stage 0: no CKM risk factors; stage 1: excess/dysfunctional adiposity; stage 2: metabolic risk factors (hypertension, diabetes, metabolic syndrome, hypertriglyceridemia) and/or moderate to high risk chronic kidney disease (CKD); stage 3: ≥ 20% predicted CVD risk using the PREVENT calculator or very high risk CKD, and stage 4: overt CVD overlapping with CKM risk factors. Among those in CKM stages 0-3 at Visit 2 (baseline), we used logistic regression to evaluate the associations of higher levels of adiponectin, leptin, and resistin at Visit 2 (modeled per 1-SD and according to tertiles) with the odds of progression to a higher CKM stage by Visit 4. Results: The mean age was 57 years (56% female, 79% White). At baseline, 6% had stage 0 CKM syndrome, 24% had stage 1 CKM syndrome, 68% had stage 2 CKM syndrome, and 2% had stage 3 CKM syndrome. Overall, 33% of participants progressed to a worse CKM stage during the 6 years of follow-up. After adjustment for sociodemographics, lifestyle factors, and CKM stage at baseline, higher adiponectin (per 1-SD) was associated with 24% lower odds of progressing to a worse CKM stage (OR 0.76, 95% CI[0.71, 0.81]). Conversely, higher leptin and resistin (per 1-SD) were associated with 29% (OR 1.29, 95% CI[1.19, 1.39]) and 9% (OR 1.09, 95% CI[1.03, 1.15] higher odds of CKM stage progression, respectively. Graded associations with similar directionality as those described were seen when baseline adipokine levels were modeled in tertiles (Table). Conclusion: Adipokines are significantly associated with CKM syndrome progression. These proteins may help to inform the prediction and prevention of a worsening trajectory of CKM syndrome over time.
Abstract P3103: Circulating and tissue levels of omega-3 fatty acids and incident peripheral artery disease: an individual participant-level pooled analysis of prospective studies
Background: Evidence from observational studies and randomized trials suggest a favorable role of omega-3 fatty acids on cardiovascular outcomes. However, whether omega-3 fatty acids may reduce the incidence of peripheral artery disease (PAD) is not known. Aims: We prospectively evaluated blood and adipose tissue levels of alpha-linolenic acid (ALA), eicosapentaenoic acid (EPA), docosapentaenoic acid (DPA), docosahexaenoic acid (DHA), and the sum of EPA and DHA, with respect to incident PAD. Methods: We included 11 prospective studies from a global consortium up to May 2023 with measurements of ALA, EPA, DPA, or DHA (as a % of total fatty acids) in blood or adipose tissue among adults (age≥18), who were free of coronary artery disease, stroke, and PAD at baseline, and assessed incident PAD events. Each cohort conducted de novo individual-level analyses with a prespecified analytical plan and harmonized definitions for exposures, outcomes, covariates (including demographics, cardiovascular risk factors, medication use, and omega-6 fatty acid levels), and subgroups. Pooled hazard ratios (HRs) were calculated using inverse-variance weighted meta-analysis. Results: Among 142,316 participants from North America, Europe, and Australia followed for a weighted median of 13.7 years (range of median follow-up: 9.8 to 26.2 years), 1,842 incident cases of PAD were ascertained. In multivariable-adjusted pooled analysis, very long-chain EPA, DPA, DHA, and EPA+DHA each associated with lower PAD incidence, with HRs (95% CI) per interquintile range of 0.79 (0.69, 0.91), 0.80 (0.68, 0.95), 0.74 (0.64, 0.85), and 0.79 (0.69, 0.90), respectively ( P <0.01 for each) ( Figure 1 ). Long-chain ALA was not associated with PAD, 1.14 (0.98, 1.33). Heterogeneity between studies was low to moderate ( I 2\ ranging from 0% to 57%). Results were broadly consistent in prespecified subgroups by age, sex, smoking status, prevalent diabetes, and biomarker lipid fraction. Conclusion: In this large international consortium, objective levels of very long-chain, but not long-chain, omega-3 fatty acids inversely associated with incident PAD. Our data suggest a protective role for very long-chain omega-3 fatty acids in the development of PAD, supporting the need for further mechanistic studies and appropriately powered randomized controlled trials.