Objective: High sedentary behavior (SB) and low physical activity (PA) can worsen glycemic control, impairing cardiac autonomic function (CAF) and increasing cardiovascular disease (CVD) risk. We examined 1) associations of PA and SB with CVD risk and 2) whether CAF, assessed by heart rate variability (HRV), mediated these associations overall and in those with and without type 2 diabetes (T2D). Methods: We analyzed data from 28,897 participants from six NHLBI cohorts: Atherosclerosis Risk in Communities Study, Coronary Artery Risk Development in Young Adults Study, Cardiovascular Health Study, Framingham Heart Study, Jackson Heart Study, and the Multi-Ethnic Study of Atherosclerosis. We harmonized self-reported PA and SB data as the cumulative average of PA and SB percentiles over all timepoints prior to a fatal/non-fatal CVD event or end of follow-up. HRV was expressed as the SD of normal-to-normal RR intervals (SDNN)—a time-domain measure of HRV and our proxy for CAF. Covariates included harmonized data on age, race/ethnicity, sex, study center, education, smoking, alcohol, diet, and lipid, blood pressure, and T2D medications. We used Cox regression to examine associations of PA and SB with CVD risk, reporting standardized betas for PA and SB. Estimate pooling was performed with a random effects model. HRV mediation was the % difference in pooled PA and SB betas when HRV was in vs. out of the model, with ≥10% beta attenuation defined as mediation. Results: Across cohorts, mean age was 30-71 years, with 8,780 CVD events observed and a mean follow-up of 7,090 days until a CVD event. PA was strongly inversely associated with CVD risk, while SB had a marginal positive association with CVD risk. There was no evidence of CAF mediation for PA. For SB, we observed no evidence of CAF mediation overall but, for those with and without T2D, we observed 16.1% and 7.7% mediation, respectively (Table). Conclusion: Additional research is needed with more robust HRV and SB measures to discern whether the relationship between SB and CVD risk is truly mediated by CAF and whether this mediation varies by T2D status.

Introduction: Many studies report physical activity (PA) and sedentary behavior (SB) to be independently associated with lower and higher cardiovascular disease (CVD) risk, respectively. We evaluated a combined exposure of PA and SB in association with CVD risk. Methods: Data were gathered from 35,844 participants from six NHLBI cohorts: Atherosclerosis Risk in Communities Study, Coronary Artery Risk Development in Young Adults Study, Cardiovascular Health Study, Framingham Heart Study, Jackson Heart Study, and the Multi-Ethnic Study of Atherosclerosis. We harmonized time-varying PA and SB into quintiles, considering the cumulative average over all timepoints prior to a first CVD event (fatal and non-fatal) or end of follow-up. We created a 6-level ‘PASB index’ based on combining the PA and SB quintiles. We performed Cox regression to examine the associations of PA and SB quintiles, and the PASB index with incident CVD risk. Covariates were age, sex, race/ethnicity, study center, education, smoking, alcohol use, and dietary quality. A random effects model was used for the pooled results. Results: Across cohorts, there were 10,913 CVD events. In pooled analyses, PA had a graded inverse association with CVD risk (HR=0.92 per quintile increment; 95% CI: 0.89, 0.95), and SB had a non-linear threshold association with CVD with the highest SB quintile having a greater risk relative to the lowest quintile (HR=1.13; 95% CI: 1.01, 1.26). Results for the six-level PASB index are shown in Figure 1, with the lowest HRs for those with low to moderate SB and either moderate or high PA. Conclusion: The combined behaviors of moderate to high PA and low SB were associated with the lowest risk for CVD. These findings support current public guidelines to engage in regular PA of at least moderate intensity and to avoid high levels of SB.

Introduction: Non-alcoholic fatty liver disease (NAFLD) is often unrecognized and untreated despite its high prevalence (~25% of adults) and implications for cardiometabolic diseases. We examined the potential for NAFLD to modify the associations between BMI and waist circumference (WC) with changes in cardiometabolic risk. Hypothesis: The associations of BMI and WC with CMR risk are modified by NAFLD. Methods: We analyzed data from Years 25 (2010/11) and 30 (2015/6) of the CARDIA cohort. NAFLD status was determined by noncontrast CT with a liver attenuation cut-off <51 HU after exclusions for other causes of liver fat. The outcome was 5-yr change in cardiometabolic risk score (CMR) derived by the mean Z-score score of fasting glucose, insulin, triglycerides [log], (-)HDL cholesterol, and systolic blood pressure. BMI and WC were directly assessed, and covariates included age, sex, race, center, education, smoking, alcohol, and medications for diabetes or CVD, and Year 25 CMR. Multivariable linear regression models included NAFLD status, BMI, and WC as the independent variables. Effect modification was assessed by interaction terms between NAFLD and continuous BMI and WC in separate models. Subsequent models were stratified by NAFLD status. Results: Our final sample size was 2366 (mean age: 50.1 y, males: 42.6%, Black race: 47.2%) with a NAFLD prevalence of 23% (n=539). Year 25 mean (± sd) CMR for those with and without NAFLD was -0.21 ± 0.58 and 0.50 ± 0.60, respectively. NAFLD status considerably modified the association of both BMI (interaction p<0.001) and WC (interaction p=0.002) with CMR change. BMI and WC were more strongly associated with CMR in participants without NAFLD compared to those with NAFLD (see Table). Within each BMI and WC category, CMR change was higher for those with NAFLD relative to those without NAFLD. Conclusion: NAFLD status modifies the association between BMI and WC with CMR changes. These findings have implications for routine clinical screening guidelines and risk stratification.

Type 2 diabetes (T2D) is a heterogeneous disease that develops through diverse pathophysiological processes1,2 and molecular mechanisms that are often specific to cell type3,4. Here, to characterize the genetic contribution to these processes across ancestry groups, we aggregate genome-wide association study data from 2,535,601 individuals (39.7% not of European ancestry), including 428,452 cases of T2D. We identify 1,289 independent association signals at genome-wide significance (P < 5 × 10-8) that map to 611 loci, of which 145 loci are, to our knowledge, previously unreported. We define eight non-overlapping clusters of T2D signals that are characterized by distinct profiles of cardiometabolic trait associations. These clusters are differentially enriched for cell-type-specific regions of open chromatin, including pancreatic islets, adipocytes, endothelial cells and enteroendocrine cells. We build cluster-specific partitioned polygenic scores5 in a further 279,552 individuals of diverse ancestry, including 30,288 cases of T2D, and test their association with T2D-related vascular outcomes. Cluster-specific partitioned polygenic scores are associated with coronary artery disease, peripheral artery disease and end-stage diabetic nephropathy across ancestry groups, highlighting the importance of obesity-related processes in the development of vascular outcomes. Our findings show the value of integrating multi-ancestry genome-wide association study data with single-cell epigenomics to disentangle the aetiological heterogeneity that drives the development and progression of T2D. This might offer a route to optimize global access to genetically informed diabetes care.

Introduction: Neighborhood green (e.g., parks) and blue (e.g., lakes) spaces have been associated with reduced stress and cardiometabolic risk (CMR), but research is limited in socially vulnerable urban communities affected by structural racism and redlining. This study explored the spatial correlation between green and blue spaces and cortisol levels, BMI percentile, waist circumference (WC), pulse, and systolic (SBP) and diastolic blood (DBP) pressure in a sample of racially/ethnically diverse children. Methods: We used cross-sectional data from the Family Matters cohort study, consisting of 1307 racially/ethnically diverse families with children (age 7±1.5 yrs). Participant addresses were matched to the US 2010 Census tract attributes. We used 246 sample locations across 149 census tracts in the 7-county Twin Cities, MN, metro area. Green and blue spaces were retrieved from the Twin Cities Metropolitan Council. Hair cortisol (chronic stress biomarker), BMI, WC, pulse, SBP and DBP were objectively measured in a subsample of 246 children. We used Decision Tree and Geographically Weighted Regression (GWR) to examine the spatially varying associations of CDC’s Social Vulnerability Index (SVI) and green and blue spaces (independent variables) with biomarkers of stress and CMR (dependent variables). SVI reflects four themes: socioeconomic status, household composition and disability, racial and ethnic minority status and language, and housing type and transportation. Results: Northwest Minneapolis was identified as a hotspot for elevated cortisol, BMI, and WC, while east St. Paul was a hotspot for elevated BMI and WC. Decision trees indicated that green and blue spaces explain variation in accumulated cortisol after tract socioeconomic status and racial and ethnic minority status. GWR revealed nonuniform relationships with the highest explanatory power in high social vulnerability hotspots. GWR’s R 2 for cortisol (0.07), BMI (0.14), and WC (0.14) were driven by hotspots where local R 2 was 2-5 times stronger than areas not identified as hotspots indicating robust geographical variations in the northern areas (areas of high social vulnerability) and distance from green and blue spaces. Patterns were inconsistent for pulse, SBP, and DBP. Conclusions: The spatially varying relationships of green and blue spaces with biomarkers of stress and CMR in children were heterogeneous, possibly indicating different mechanisms across biomarkers and geographical areas. Higher cortisol levels, BMI percentile, and WC in northern Twin Cities areas may reflect the legacy of structural racism and redlining, which have resulted in inequitable use of green and blue spaces for minoritized communities. Future interventions co-developed with local communities to improve green and blue spaces’ access and use, as well as to reduce disparities in stress and CMR, are needed to advance health equity.

Introduction: Cardiovascular diseases (CVDs) account for ~33% of all deaths globally. Aerobic exercise (AE) is a well-established modifiable risk factor for lowering the risk of CVD mortality, but the global prevalence of meeting the World Health Organization’s (WHO’s) recommendation for weekly AE in adults is only ~25%. Given the rapidly growing body of evidence indicating resistance exercise (RE) confers numerous health benefits independent of AE, and may also lower the risk of CVD mortality, we conducted a living meta-analysis (PROSPERO ID: CRD42022356528) to determine: 1) if RE is associated with lower CVD mortality risk independent of, and jointly with, AE; and 2) whether there is a dose-response pattern of CVD mortality by amount of RE. Methods: We searched MEDLINE, Embase, SPORTDiscus, The Cochrane Library, and SCOPUS from inception to October 10, 2023. We included prospective cohort studies in which: 1) AE and RE were the exposures and CVD mortality the outcome; and 2) nonclinical adult populations (aged ≥18 years) were sampled. Study-specific hazard ratios (HRs) and 95% confidence intervals (CIs) for CVD mortality were pooled using random effects meta-analysis models. The independent effect of RE from AE was based on models with AE as a covariate. A random effects dose-response meta-analysis characterized the association between the RE dose (sessions/week) and CVD mortality risk. Results: Ten cohort studies from 5 countries (Mean follow-up time = 10.1 years) were included (9, 8, and 5 examined independent, joint, and dose-response associations, respectively). Relative to no RE, meeting the WHO guideline for RE (1-2 sessions/week), independent of meeting the AE guideline (≥150 minutes/week at moderate to vigorous intensities), was associated with a lower risk of CVD mortality (HR: 0.90, 95% CI: 0.85-0.95). A sensitivity analysis removing a highly weighted study (53%) yielded consistent results. Compared to meeting neither guideline, meeting the joint AE-RE guideline was associated with a larger difference in CVD mortality risk (HR: 0.60, 95% CI: 0.45-0.81). The dose-response meta-analysis showed evidence of a non-linear, U-shaped relationship between the amount of RE (independent of AE) and risk of CVD mortality. In detail, relative to no RE, participating in 1-2 sessions/week of RE was associated with a lower risk of CVD mortality (HR: 0.81; 95% CI: 0.72, 0.91), but there was no difference in CVD mortality risk between those who exceeded the guideline and those who participated in no RE (HR: 1.00, 95% CI: 0.89, 1.12). Conclusion: With the increasing global prevalence of obesity and physical inactivity, RE—a more movement-varied, lower-impact exercise modality—should receive more attention in public health policy to help alleviate the global burden of CVD mortality. Public health surveillance instruments capturing RE volume (sets, repetitions) are needed to improve prescription.

ABSTRACT Habitual physical activity (PA) impacts the plasma proteome and reduces the risk of developing type 2 diabetes (T2D). Using a large-scale proteome-wide approach in Atherosclerosis Risk in Communities Study participants, we aimed to identify plasma proteins associated with PA and determine which of these may be causally related to lower T2D risk. PA was associated with 92 plasma proteins in discovery (p<1.01x10-5), and 40 remained significant in replication (p<5.43x10-4). Eighteen of these proteins were independently associated with incident T2D (p<1.25x10-3) including neuronal growth regulator 1 (NeGR1) (HR per SD: 0.85; p-value=7.5x10-11). Two-sample Mendelian randomization inverse variance weighted analysis indicated that higher NeGR1 reduces T2D risk (OR per SD: 0.92; p=0.03) and was consistent with MR-Egger, weighted median, and weighted mode sensitivity analyses. A stronger association was observed for the single cis-acting NeGR1 genetic variant (OR per SD: 0.80; p=6.3x10-5). Coupled with previous evidence that low circulating NeGR1 levels promote adiposity, its association with PA and potential causal role in T2D shown here suggest that NeGR1 may link PA exposure with metabolic outcomes. Further research is warranted to confirm our findings and examine the interplay of PA, NeGR1, adiposity, and metabolic health. Article Highlights Physical activity alters tissue and plasma proteomes, which have been shown to have roles in type 2 diabetes development We aimed to identify protein signatures through which physical activity may influence type 2 diabetes pathogenesis Forty proteins associated with physical activity, and 18 of these were further related to incident type 2 diabetes over an approximate 24-year follow-up Two-sample Mendelian randomization analysis indicated that circulating neuronal growth regulator 1 reduces risk of type 2 diabetes · These findings suggest that plasma neuronal growth regulator 1 may link physical activity to reduced risk of type 2 diabetes

Background: Low-density lipoprotein (LDL) cholesterol change with consumption of a low-carbohydrate diet (LCD) is highly variable. Identifying the source of this heterogeneity could guide clinical decision-making. Objectives: To evaluate LDL cholesterol change in randomized controlled trials involving LCDs, with a focus on body mass index (BMI) in kg/m2. Methods: Three electronic indexes (Pubmed, EBSCO, and Scielo) were searched for studies between 1 January, 2003 and 20 December, 2022. Two independent reviewers identified randomized controlled trials involving adults consuming <130 g/d carbohydrate and reporting BMI and LDL cholesterol change or equivalent data. Two investigators extracted relevant data, which were validated by other investigators. Data were analyzed using a randomeffects model and contrasted with results of pooled individual participant data. Results: Forty-one trials with 1379 participants and a mean intervention duration of 19.4 wk were included. In a meta-regression accounting for 51.4% of the observed variability on LCDs, mean baseline BMI had a strong inverse association with LDL cholesterol change [beta = -2.5 mg/dL/BMI unit, 95% confidence interval (CI): -3.7, -1.4], whereas saturated fat amount was not significantly associated with LDL cholesterol change. For trials with mean baseline BMI <25, LDL cholesterol increased by 41 mg/dL (95% CI: 19.6, 63.3) on the LCD. By contrast, for trials with a mean of BMI 25-<35, LDL cholesterol did not change, and for trials with a mean BMI >= 35, LDL cholesterol decreased by 7 mg/dL (95% CI: -12.1, -1.3). Using individual participant data, the relationship between BMI and LDL cholesterol change was not observed on higher-carbohydrate diets. Conclusions: A substantial increase in LDL cholesterol is likely for individuals with low but not high BMI with consumption of an LCD, findings that may help guide individualized nutritional management of cardiovascular disease risk. As carbohydrate restriction tends to improve other lipid and nonlipid risk factors, the clinical significance of isolated LDL cholesterol elevation in this context warrants investigation. This trial was registered at PROSPERO as CRD42022299278.

An in fl uential 2-wk cross-over feeding trial without a washout period purported to show advantages of a low-fat diet (LFD) compared with a low-carbohydrate diet (LCD) for weight control. In contrast to several other macronutrient trials, the diet order effect was originally reported as not signi fi cant. In light of a new analysis by the original investigative group identifying an order effect, we aimed to examine, in a reanalysis of publicly available data (16 of 20 original participants; 7 female; mean BMI, 27.8 kg/m 2 ), the validity of the original results and the claims that trial data oppose the carbohydrate - insulin model of obesity (CIM). We found that energy intake on the LCD was much lower when this diet was consumed fi rst compared with second (a difference of - 1164 kcal/d, P = 3.6 x 10 -13 ); the opposite pattern was observed for the LFD (924 kcal/d, P = 2.0 x 10 -16 ). This carry-over effect was signi fi cant ( P interaction = 0.0004) whereas the net dietary effect was not ( P = 0.4). Likewise, the between-arm difference (LCD - LFD) was - 320 kcal/d in the fi rst period and + 1771 kcal/d in the second. Body fat decreased with consumption of the LCD fi rst and increased with consumption of this diet second ( - 0.69 + 0.33 compared with 0.57 + 0.32 kg, P = 0.007). LCD- fi rst participants had higher beta -hydroxybutyrate levels while consuming the LCD and lower respiratory quotients while consuming LFD when compared with LFD- fi rst participants on their respective diets. Change in insulin secretion as assessed by Cpeptide in the fi rst diet period predicted higher energy intake and less fat loss in the second period. These fi ndings, which tend to support rather than oppose the CIM, suggest that differential (unequal) carry-over effects and short duration, with no washout period, preclude causal inferences regarding chronic macronutrient effects from this trial.

This study explored differences in motivations and barriers for green physical activity (GPA) in a sample of U.S. adults (N = 205). Differences by identity (i.e. gender, sexual identity, disability status, and race/ethnicity) were examined. Participants (33.5 +/- 9.2 years, 76.6% female and 7.8% genderqueer/non-conforming, 33.3% LGBTQIA+, 34.1% BIPOC, 10.7% individuals with disabilities) responded to survey measures for motivations/barriers and could provide open-ended motivations/barriers not covered by scales. Genderqueer/non-conforming participants were significantly less motivated by competition/ego, appearance, and others' expectations (p's < 0.05). LGBTQIA + participants were overall less motivated to do GPA (p = 0.02). Overall barrier scores were significantly higher for LGBTQIA + participants (p = 0.01) and participants with disabilities (p = 0.02). Qualitative analysis revealed themes outside of the scales for motivators (i.e. GPA integrated into daily life, active citizenship, mental health, and physical activity benefits) and further barriers at all levels (i.e. structural, interpersonal, and intrapersonal). Historically excluded groups' motivators/barriers should be prioritized in GPA-related campaigns and programming.

BACKGROUND:There are few resources available for researchers aiming to conduct 24-h dietary record and recall analysis using R. OBJECTIVES:We aimed to develop DietDiveR, which is a toolkit of functions written in R for the analysis of recall or record data collected with the Automated Self-Administered 24-h Dietary Assessment Tool or 2-d 24-h dietary recall data from the National Health and Nutrition Examination Survey (NHANES). The R functions are intended for food and nutrition researchers who are not computational experts. METHODS:DietDiveR provides users with functions to 1) clean dietary data, 2) analyze 24-h dietary intakes in relation to other study-specific metadata variables, 3) visualize percentages of energy intake from macronutrients, 4) perform principal component analysis or k-means clustering to group participants by similar data-driven dietary patterns, 5) generate foodtrees based on the hierarchical food group information for food items consumed, 6) perform principal coordinate analysis taking food grouping information into account, and 7) calculate diversity metrics for overall diet and specific food groups. DietDiveR includes a self-paced tutorial on a website (https://computational-nutrition-lab.github.io/DietDiveR/). As a demonstration, we applied DietDiveR to a demonstration data set and data from NHANES 2015-2016 to derive a dietary diversity measure of nuts, seeds, and legumes consumption. RESULTS:Adult participants in the NHANES 2015-2016 cycle were grouped depending on the diversity in their mean consumption of nuts, seeds, and legumes. The group with the highest diversity in nuts, seeds, and legumes consumption had 3.8 cm lower waist circumference (95% confidence interval: 1.0, 6.5) than those who did not consume nuts, seeds, and legumes. CONCLUSIONS:DietDiveR enables users to visualize dietary data and conduct data-driven dietary pattern analyses using R to answer research questions regarding diet. As a demonstration of this toolkit, we explored the diversity of nuts, seeds, and legumes consumption to highlight some of the ways DietDiveR can be used for analyses of dietary diversity.

Introduction: Non-alcoholic fatty liver disease (NAFLD) is one of the most common causes of chronic liver disease in the US. Weight loss and lifestyle modifications, including increases in physical activity (PA), are the primary therapeutic recommendations for managing NAFLD. However, little is known about longitudinal physical activity patterns and NAFLD prevalence among a racially diverse population of adults. This study aimed to (1) identify distinct physical activity trajectories from young to middle adulthood and (2) examine the associations between physical activity trajectories and NAFLD prevalence in middle-aged adults. Methods: The analytic sample included 2833 participants of the Coronary Artery Risk Development in Young Adults (CARDIA) study. Physical activity was self-reported at eight examinations over 25 years (1985/1986 to 2010/2011), and separately scored for moderate- (MPA) and vigorous-intensity (VPA) physical activity. NAFLD was defined as liver attenuation values ≤51 Hounsfield units after exclusion of other causes of liver fat, measured using computed tomography in year 25 (2010/2011). Group-based trajectory modeling was used to identify MPA and VPA trajectories over 25 years. Modified Poisson regression models were used to estimate the risk ratios (RR) of NAFLD across the PA trajectories, adjusted for year 25 sex, age, CARDIA study center, race, education, smoking status, dietary pattern, and alcohol consumption. Results: We identified three distinct MPA and VPA trajectories. MPA: ‘very low stable’ (53%), ‘low stable’ (39%), and ‘moderate increasing’ (8%). VPA: ‘low stable’ (57%), ‘moderate stable’ (36%), and ‘high decreasing’ (7%). Year 25 NAFLD prevalence was 24%. Relative to participants with ‘moderate increasing’ MPA over time, those with lower levels of MPA had adjusted risk ratios of 1.11 (95% CI 0.87-1.43) for ‘low stable’ and 1.11 (95% CI 0.86-1.42) for ‘very low stable.’ Relative to participants with ‘high decreasing’ VPA over time, those with lower levels of VPA had adjusted risk ratios of 1.26 (95% CI 0.95-1.69) for ‘moderate stable’ and 1.62 (95% CI 1.21-2.16) for ‘low stable.’ Conclusions: Our results suggest a higher risk of NAFLD in middle age for those with low levels of VPA from young to middle adulthood relative to those with the highest levels of VPA over time, with evidence of dose-response. Sustaining VPA throughout adulthood may be efficacious in the prevention of NAFLD, highlighting the importance of targeted delivery of prevention and management programs aimed at modifying lifestyle risk factors to reduce NAFLD risk.

Background: The prevalence of non-alcoholic fatty liver disease (NAFLD) in U.S. adults is over 30%, yet the role of lifestyle factors in the etiology of NAFLD remains understudied. We examined the associations of physical activity, by intensity and type, and television viewing with prevalent NAFLD. Methods: Cross-sectional analysis of a population-based sample of 2726 Black (49%) and White (51%) adults (Mean (SD) age, 50 (3.6) years; 57.3% female) from the CARDIA study. Exposures were aerobic activity by intensity (moderate, vigorous; hours/week); activity type (aerobic, muscle-strengthening; hours/week); and television viewing (hours/week), examined concurrently in all models and assessed by validated questionnaires. Our outcome was NAFLD (liver attenuation < 51 Hounsfield Units), measured by non-contrast computed tomography, after exclusions for other causes of liver fat. Covariates were sex, age, race, study center, education, diet quality, smoking status, alcohol consumption, and body mass index or waist circumference. Results: 648 participants had NAFLD. In the fully adjusted modified Poisson regression model, the risk ratios per interquartile range of each exposure were moderate-intensity aerobic activity, 1.10 (95% CI, 0.97–1.26); vigorous-intensity aerobic activity, 0.72 (0.63–0.82); muscle-strengthening activity, 0.89 (0.80–1.01); and television viewing, 1.20 (1.10–1.32). Relative to less active participants with higher levels of television viewing, those who participated in ≥2 h/week of both vigorous-intensity aerobic and muscle-strengthening activity and <7 h/week of television viewing had 65% lower risk of NAFLD (risk ratio = 0.35, 95% CI = 0.23–0.51). Conclusion: Adults who follow public health recommendations for vigorous-aerobic and muscle-strengthening activity, as well as minimize television viewing, are considerably less likely to have NAFLD than those who do not follow the recommendations and who have relatively high levels of television viewing.

PURPOSE: To identify plasma proteins associated with habitual physical activity (PA) in a large, bi-racial cohort study of U.S. adults and estimate associations between proteins and risk of incident type 2 diabetes (T2D). METHODS: Data were from visit 3 (1993-95) of the community-based Atherosclerosis Risk in Communities (ARIC) study (N = 10,644; X̅age ± SD = 60 ± 6 years; 54 % female; 21% Black). Habitual PA (past-year sport, leisure, and work domains) was assessed by the validated Baecke questionnaire, analyzed as an index score from 1 (low) to 5 (high) based on self-reported frequency, duration, and intensity. Plasma protein concentrations were quantified using a multiplexed modified DNA-based aptamer assay. Cross-sectional PA-protein associations were estimated using multivariable linear regression in a randomly selected discovery sample (2/3 N = 7,093), with Bonferroni correction for multiple comparisons. We adjusted for traditional covariates in addition to prevalent coronary heart disease/stroke, estimated glomerular filtration rate, BMI, and waist:hip ratio. We confirmed statistically significant protein-PA associations in a replication sample of 3,551 participants. After exclusion for prevalent T2D, we evaluated replicated proteins for their association with incident T2D (2,259 events in 6,039 participants) over a median 24-year follow up. Decided a priori, the 10 proteins most strongly associated with T2D were interrogated using a two-sample Mendelian randomization (MR) analysis. RESULTS: In discovery, PA was associated with 92 plasma proteins (p < 1.01 x 10-5); 40 remained significant in replication (p < 5.43 x 10-4). Of these proteins, 18 were independently associated with incident T2D (p < 1.25 x 10-3), most notably neuronal growth regulator 1 (NeGR1; hazard ratio per SD: 0.85; p = 7.5 x 10-11). MR inverse variance weighted analysis indicated that higher plasma levels of NeGR1 were related to lower risk of T2D (odds ratio [OR] per SD: 0.92; p = 0.03), consistent with MR-Egger, weighted median, and weighted mode sensitivity analyses (not shown). A stronger association was observed for the single cis-acting NeGR1 SNP (OR per SD: 0.80; p = 6.3 x 10-5). CONCLUSION: Low circulating NeGR1 requires further study as a potential novel biomarker for incident T2D and a target for mechanistic PA interventions.