Apt interpretation of comprehensive lipoprotein data in large-scale epidemiology: disclosure of fundamental structural and metabolic relationships

Fasting and postprandial plasma metabolite responses to a 12-week dietary intervention in tissue-specific insulin resistance: a secondary analysis of the PERSON randomized trial

BACKGROUND

We previously showed that dietary intervention effects on cardiometabolic health were driven by tissue-specific insulin resistance (IR) phenotype: individuals with predominant muscle IR (MIR) benefitted more from a low-fat, high-protein, high-fiber diet (LFHP), while individuals with predominant liver IR (LIR) benefitted more from a diet rich in mono-unsaturated fat (HMUFA).

OBJECTIVE

To further characterize the effects of LFHP and HMUFA diets and their interaction with tissue-specific IR, we investigated dietary intervention effects on fasting and postprandial plasma metabolite profile.

METHODS

Adults with MIR or LIR (40-75 years, BMI 25-40 kg/m2) were randomized to a 12-week HMUFA or LFHP diet (n=242). After exclusion of statin use, 214 participants were included in this pre-specified secondary analysis. Plasma samples were collected before (T=0) and after (T=30, 60, 120, 240 min) a high-fat mixed meal for quantification of 247 metabolite measures using nuclear magnetic resonance spectroscopy.

RESULTS

A larger reduction in fasting VLDL-TAG and VLDL particle size was observed in individuals with MIR following the LFHP diet and those with LIR following the HMUFA diet, although no longer statistically significant after false discovery rate (FDR) adjustment. No IR phenotype-diet interactions were found for postprandial plasma metabolites assessed as total area under the curve (tAUC). Irrespective of IR phenotype, the LFHP diet induced greater reductions in postprandial plasma tAUC of the larger VLDL particles and small HDL particles, and TAG content in most VLDL subclasses and the smaller LDL and HDL subclasses (e.g. VLDL-TAG tAUC standardized mean change [95% CI] LFHP = -0.29 [-0.43, -0.16] compared to HMUFA = -0.04 [-0.16, 0.09]; FDR-adjusted P for Diet x Time = 0.041).

CONCLUSIONS

Diet effects on plasma metabolite profiles were more pronounced than phenotype-diet interactions. A LFHP diet may be more effective than a HMUFA diet for reducing cardiometabolic risk in individuals with tissue-specific IR, irrespective of IR phenotype.

CLINICALTRIALS

GOV REGISTRATION

NCT03708419, https://clinicaltrials.gov/study/NCT03708419?term=NCT03708419&rank=1 CCMO REGISTRATION: NL63768.068.17, https://www.toetsingonline.nl/to/ccmo_search.nsf/fABRpop?readform&unids=3969AABCD9BA27FEC12587F1001BCC65.

Blood Lipids, Lipoproteins, and Apolipoproteins With Risk of Coronary Heart Disease: A Prospective Study Among Racially Diverse Populations

BACKGROUND

Comprehensive blood lipoprotein profiles and their association with incident coronary heart disease (CHD) among racially and geographically diverse populations remain understudied.

METHODS AND RESULTS

We conducted nested case-control studies of CHD among 3438 individuals (1719 pairs), including 1084 White Americans (542 pairs), 1244 Black Americans (622 pairs), and 1110 Chinese adults (555 pairs). We examined 36 plasma lipids, lipoproteins, and apolipoproteins, measured by nuclear magnetic resonance spectroscopy, with incident CHD among all participants and subgroups by demographics, lifestyle, and metabolic health status using conditional or unconditional logistic regression adjusted for potential confounders. Conventionally measured blood lipids, that is, total cholesterol, triglycerides, low-density lipoprotein-cholesterol, and high-density lipoprotein-cholesterol, were each associated with incident CHD, with odds ratios (ORs) being 1.33, 1.32, 1.24, and 0.79 per 1-SD increase among all participants. Seventeen lipoprotein biomarkers showed numerically stronger associations than conventional lipids, with ORs per 1-SD among all participants ranging from 1.35 to 1.57 and a negative OR of 0.78 (all false discovery rate <0.05), including apolipoprotein B100 to apolipoprotein A1 ratio (OR, 1.57 [95% CI, 1.45-1.7]), low-density lipoprotein-triglycerides (OR, 1.55 [95% CI, 1.43-1.69]), and apolipoprotein B (OR, 1.49 [95% CI, 1.37-1.62]). All these associations were significant and consistent across racial groups and other subgroups defined by age, sex, smoking, obesity, and metabolic health status, including individuals with normal levels of conventionally measured lipids.

CONCLUSIONS

Our study highlighted several lipoprotein biomarkers, including apolipoprotein B/ apolipoprotein A1 ratio, apolipoprotein B, and low-density lipoprotein-triglycerides, strongly and consistently associated with incident CHD. Our results suggest that comprehensive lipoprotein measures may complement the standard lipid panel to inform CHD risk among diverse populations.

Gene expression profiling in elderly patients with familial hypercholesterolemia with and without coronary heart disease

BACKGROUND AND AIMS

Elderly familial hypercholesterolemia (FH) patients are at high risk of coronary heart disease (CHD) due to high cholesterol burden and late onset of effective cholesterol-lowering therapies. A subset of these individuals remains free from any CHD event, indicating the potential presence of protective factors. Identifying possible cardioprotective gene expression profiles could contribute to our understanding of CHD prevention and future preventive treatment. Therefore, this study aimed to investigate gene expression profiles in elderly event-free FH patients.

METHODS

Expression of 773 genes was analysed using the Nanostring Metabolic Pathways Panel, in peripheral blood mononuclear cells (PBMCs) from FH patients ≥65 years without CHD (FH event-free, n = 44) and with CHD (FH CHD, n = 39), and from healthy controls ≥70 years (n = 39).

RESULTS

None of the genes were differentially expressed between FH patients with and without CHD after adjusting for multiple testing. However, at nominal p < 0.05, we found 36 (5%) differentially expressed genes (DEGs) between the two FH groups, mainly related to lipid metabolism (e.g. higher expression of ABCA1 and ABCG1 in FH event-free) and immune responses (e.g. lower expression of STAT1 and STAT3 in FH event-free). When comparing FH patients to controls, the event-free group had fewer DEGs than the CHD group; 147 (19%) and 219 (28%) DEGs, respectively.

CONCLUSIONS

Elderly event-free FH patients displayed a different PBMC gene expression profile compared to FH patients with CHD. Differences in gene expression compared to healthy controls were more pronounced in the CHD group, indicating a less atherogenic gene expression profile in event-free individuals. Overall, identification of cardioprotective factors could lead to future therapeutic targets.

Clinical and biochemical associations of urinary metabolites: quantitative epidemiological approach on renal-cardiometabolic biomarkers

BACKGROUND

Urinary metabolomics has demonstrated considerable potential to assess kidney function and its metabolic corollaries in health and disease. However, applications in epidemiology remain sparse due to technical challenges.

METHODS

We added 17 metabolites to an open-access urinary nuclear magnetic resonance metabolomics platform, extending the panel to 61 metabolites (n = 994). We also introduced automated quantification for 11 metabolites, extending the panel to 12 metabolites (+creatinine). Epidemiological associations between these 12 metabolites and 49 clinical measures were studied in three independent cohorts (up to 5989 participants). Detailed regression analyses with various confounding factors are presented for body mass index (BMI) and smoking.

RESULTS

Sex-specific population reference concentrations and distributions are provided for 61 urinary metabolites (419 men and 575 women), together with methodological intra-assay metabolite variations as well as the biological intra-individual and epidemiological population variations. For the 12 metabolites, 362 associations were found. These are mostly novel and reflect potential molecular proxies to estimate kidney function, as the associations cannot be simply explained by estimated glomerular filtration rate. Unspecific renal excretion results in leakage of amino acids (and glucose) to urine in all individuals. Seven urinary metabolites associated with smoking, providing questionnaire-independent proxy measures of smoking status in epidemiological studies. Common confounders did not affect metabolite associations with smoking, but insulin had a clear effect on most associations with BMI, including strong effects on 2-hydroxyisobutyrate, valine, alanine, trigonelline and hippurate.

CONCLUSIONS

Urinary metabolomics provides new insight on kidney function and related biomarkers on the renal-cardiometabolic system, supporting large-scale applications in epidemiology.

Short-Term Metabolic Changes and Their Physiological Mediators in the Roux-en-Y Gastric Bypass Bariatric Surgery

BACKGROUND

The Roux-en-Y gastric bypass (RYGB) is a common bariatric surgery to treat obesity. Its metabolic consequences are favourable and long-term clinical corollaries beneficial. However, detailed assessments of various affected metabolic pathways and their mediating physiological factors are scarce.

METHODS

We performed a clinical study with 30 RYGB patients in preoperative and 6-month postoperative visits. NMR metabolomics was applied to profiling of systemic metabolism via 80 molecular traits, representing core cardiometabolic pathways. Glucose, glycated haemoglobin (HbA1c), insulin, and apolipoprotein B-48 were measured with standard assays. Logistic regression models of the surgery effect were used for each metabolic measure and assessed individually for multiple mediating physiological factors.

RESULTS

Changes in insulin concentrations reflected those of BMI with robust decreases due to the surgery. Six months after the surgery, triglycerides, remnant cholesterol, and apolipoprotein B-100 were decreased -24%, -18%, and -14%, respectively. Lactate and glycoprotein acetyls, a systemic inflammation biomarker, decreased -16% and -9%, respectively. The concentrations of branched-chain (BCAA; leucine, isoleucine, and valine) and aromatic (phenylalanine and tyrosine) amino acids decreased after the surgery between -17% for tyrosine and -23% for leucine. Except for the most prominent metabolic changes observed for the BCAAs, all changes were almost completely mediated by weight change and insulin. Glucose and type 2 diabetes had clearly weaker effects on the metabolic changes.

CONCLUSIONS

The comprehensive metabolic analyses indicate that weight loss and improved insulin sensitivity during the 6 months after the RYGB surgery are the key physiological outcomes mediating the short-term advantageous metabolic effects of RYGB. The clinical study was registered at ClinicalTrials.gov as NCT01330251.

Association of tobacco smoke exposure with metabolic profile from childhood to early adulthood: the Special Turku Coronary Risk Factor Intervention Project

AIMS

To investigate the associations between passive tobacco smoke exposure and daily smoking with a comprehensive metabolic profile, measured repeatedly from childhood to adulthood.

METHODS AND RESULTS

Study cohort was derived from the Special Turku Coronary Risk Factor Intervention Project (STRIP). Smoking status was obtained by questionnaire, while serum cotinine concentrations were measured using gas chromatography. Metabolic measures were quantified by nuclear magnetic resonance metabolomics at 9 (n = 539), 11 (n = 536), 13 (n = 525), 15 (n = 488), 17 (n = 455), and 19 (n = 409) years. Association of passive tobacco smoke exposure with metabolic profile compared participants who reported less-than-weekly smoking and had serum cotinine concentration <1 ng/mL (no exposure) with those whose cotinine concentration was ≥10 ng/mL (passive tobacco smoke exposure). Associations of daily smoking with metabolic profile in adolescence were analysed by comparing participants reporting daily smoking with those reporting no tobacco use and having serum cotinine concentrations <1 ng/mL. Passive tobacco smoke exposure was directly associated with the serum ratio of monounsaturated fatty acids to total fatty acids [β = 0.34 standard deviation (SD), (0.17-0.51), P < 0.0001] and inversely associated with the serum ratios of polyunsaturated fatty acids. Exposure to passive tobacco smoke was directly associated with very-low-density lipoprotein particle size [β = 0.28 SD, (0.12-0.45), P = 0.001] and inversely associated with HDL particle size {β = -0.21 SD, [-0.34 to -0.07], P = 0.003}. Daily smokers exhibited a similar metabolic profile to those exposed to passive tobacco smoke. These results persisted after adjusting for body mass index, STRIP study group allocation, dietary target score, pubertal status, and parental socio-economic status.

CONCLUSION

Both passive and active tobacco smoke exposures during childhood and adolescence are detrimentally associated with circulating metabolic measures indicative of increased cardio-metabolic risk.

HDL Particle Concentration and Size Predict Incident Coronary Artery Disease Events in People with Type 1 Diabetes

Background

Cholesterol efflux capacity (CEC) negatively correlates with cardiovascular disease risk. Small HDL particles account almost quantitively for CEC, perhaps mediated through efflux of outer leaflet plasma membrane phospholipids by ABCA1. People with type 1 diabetes (T1D) are at increased risk of coronary artery disease (CAD) despite normal levels of HDL-cholesterol (HDL-C). We therefore tested the hypotheses that small HDL particles (HDL-P)-rather than HDL-C levels-predict incident CAD in T1D.

Methods

Incident CAD (CAD death, myocardial infarction, and/or coronary revascularization) was determined in a cohort of 550 participants with childhood-onset T1D. HDL-P was quantified by calibrated ion mobility analysis. CEC and phospholipid efflux were quantified with validated assays.

Results

During a median follow-up of 26 years, 36.5% of the participants developed incident CAD. In multivariable Cox models, levels of HDL-C and apolipoprotein A-I (APOA1) did not predict CAD risk. In contrast, extra-small HDL particle levels strongly and negatively predicted risk (hazard ratio [HR]=0.25, 95% confidence interval [CI]=0.13-0.49). An increased concentration of total HDL particles (T-HDL-P) (HR=0.87, CI=0.82-0.92) and three other HDL sizes were weaker predictors of risk: small HDL (HR=0.80, 0.65-0.98), medium HDL (HR=0.78, CI=0.70-0.87) and large HDL (HR=0.72, CI=0.59-0.89). Although CEC negatively associated with incident CAD, that association disappeared after the model was adjusted for T-HDL-P. Isolated small HDLs strongly promoted ABCA1-dependent efflux of membrane outer leaflet phospholipids.

Conclusions

Low concentrations of T-HDL-P and all four sizes of HDL subpopulations predicted incident CAD independently of HDL-C, APOA1, and other common CVD risk factors. Extra-small HDL was a much stronger predictor of risk than the other HDLs. Our data are consistent with the proposal that small HDLs play a critical role in cardioprotection in T1D, which might be mediated by macrophage plasma membrane outer leaflet phospholipid export and cholesterol efflux by the ABCA1 pathway.

Fingerprinting and profiling in metabolomics of biosamples

This review focuses on metabolomics from an NMR point of view. It attempts to cover the broad scope of metabolomics and describes the NMR experiments that are most suitable for each sample type. It is addressed not only to NMR specialists, but to all researchers who wish to approach metabolomics with a clear idea of what they wish to achieve but not necessarily with a deep knowledge of NMR. For this reason, some technical parts may seem a bit naïve to the experts. The review starts by describing standard metabolomics procedures, which imply the use of a dedicated 600 MHz instrument and of four properly standardized 1D experiments. Standardization is a must if one wants to directly compare NMR results obtained in different labs. A brief mention is also made of standardized pre-analytical procedures, which are even more essential. Attention is paid to the distinction between fingerprinting and profiling, and the advantages and disadvantages of fingerprinting are clarified. This aspect is often not fully appreciated. Then profiling, and the associated problems of signal assignment and quantitation, are discussed. We also describe less conventional approaches, such as the use of different magnetic fields, the use of signal enhancement techniques to increase sensitivity, and the potential of field-shuttling NMR. A few examples of biomedical applications are also given, again with the focus on NMR techniques that are most suitable to achieve each particular goal, including a description of the most common heteronuclear experiments. Finally, the growing applications of metabolomics to foodstuffs are described.

Low concentrations of medium-sized HDL particles predict incident CVD in chronic kidney disease patients

Patients with chronic kidney disease (CKD) are at high risk for CVD. However, traditional CVD risk factors cannot completely explain the increased risk. Altered HDL proteome is linked with incident CVD in CKD patients, but it is unclear whether other HDL metrics are associated with incident CVD in this population. In the current study, we analyzed samples from two independent prospective case-control cohorts of CKD patients, the Clinical Phenotyping and Resource Biobank Core (CPROBE) and the Chronic Renal Insufficiency Cohort (CRIC). We measured HDL particle sizes and concentrations (HDL-P) by calibrated ion mobility analysis and HDL cholesterol efflux capacity (CEC) by cAMP-stimulated J774 macrophages in 92 subjects from the CPROBE cohort (46 CVD and 46 controls) and in 91 subjects from the CRIC cohort (34 CVD and 57 controls). We tested associations of HDL metrics with incident CVD using logistic regression analysis. No significant associations were found for HDL-C or HDL-CEC in either cohort. Total HDL-P was only negatively associated with incident CVD in the CRIC cohort in unadjusted analysis. Among the six sized HDL subspecies, only medium-sized HDL-P was significantly and negatively associated with incident CVD in both cohorts after adjusting for clinical confounders and lipid risk factors with odds ratios (per 1-SD) of 0.45 (0.22-0.93, P = 0.032) and 0.42 (0.20-0.87, P = 0.019) for CPROBE and CRIC cohorts, respectively. Our observations indicate that medium-sized HDL-P-but not other-sized HDL-P or total HDL-P, HDL-C, or HDL-CEC-may be a prognostic cardiovascular risk marker in CKD.

Interactions between the lipidome and genetic and environmental factors in autism

Autism omics research has historically been reductionist and diagnosis centric, with little attention paid to common co-occurring conditions (for example, sleep and feeding disorders) and the complex interplay between molecular profiles and neurodevelopment, genetics, environmental factors and health. Here we explored the plasma lipidome (783 lipid species) in 765 children (485 diagnosed with autism spectrum disorder (ASD)) within the Australian Autism Biobank. We identified lipids associated with ASD diagnosis (n = 8), sleep disturbances (n = 20) and cognitive function (n = 8) and found that long-chain polyunsaturated fatty acids may causally contribute to sleep disturbances mediated by the FADS gene cluster. We explored the interplay of environmental factors with neurodevelopment and the lipidome, finding that sleep disturbances and unhealthy diet have a convergent lipidome profile (with potential mediation by the microbiome) that is also independently associated with poorer adaptive function. In contrast, ASD lipidome differences were accounted for by dietary differences and sleep disturbances. We identified a large chr19p13.2 copy number variant genetic deletion spanning the LDLR gene and two high-confidence ASD genes (ELAVL3 and SMARCA4) in one child with an ASD diagnosis and widespread low-density lipoprotein-related lipidome derangements. Lipidomics captures the complexity of neurodevelopment, as well as the biological effects of conditions that commonly affect quality of life among autistic people.

Association of physical activity with metabolic profile from adolescence to adulthood

OBJECTIVE

Physical activity benefits cardiometabolic health, but little is known about its detailed links with serum lipoproteins, amino acids, and glucose metabolism at young age. We therefore studied the association of physical activity with a comprehensive metabolic profile measured repeatedly in adolescence.

METHODS

The cohort is derived from the longitudinal Special Turku Coronary Risk Factor Intervention Project. At ages 13, 15, 17, and 19 years, data on physical activity were collected by a questionnaire, and circulating metabolic measures were quantified by nuclear magnetic resonance metabolomics from repeatedly assessed serum samples (age 13: n = 503, 15: n = 472, 17: n = 466, and 19: n = 361).

RESULTS

Leisure-time physical activity (LTPA;MET h/wk) was directly associated with concentrations of polyunsaturated fatty acids, and inversely with the ratio of monounsaturated fatty acids to total fatty acids (-0.006SD; [-0.008, -0.003]; p < 0.0001). LTPA was inversely associated with very-low-density lipoprotein (VLDL) particle concentration (-0.003SD; [-0.005, -0.001]; p = 0.002) and VLDL particle size (-0.005SD; [-0.007, -0.003]; p < 0.0001). LTPA showed direct association with the particle concentration and size of high-density lipoprotein (HDL), and HDL cholesterol concentration (0.004SD; [0.002, 0.006]; p < 0.0001). Inverse associations of LTPA with triglyceride and total lipid concentrations in large to small sized VLDL subclasses were found. Weaker associations were seen for other metabolic measures including inverse associations with concentrations of lactate, isoleucine, glycoprotein acetylation, and a direct association with creatinine concentration. The results remained after adjusting for body mass index and proportions of energy intakes from macronutrients.

CONCLUSIONS

Physical activity during adolescence is beneficially associated with the metabolic profile including novel markers. The results support recommendations on physical activity during adolescence to promote health and possibly reduce future disease risks.

Quality control and removal of technical variation of NMR metabolic biomarker data in ~120,000 UK Biobank participants

Metabolic biomarker data quantified by nuclear magnetic resonance (NMR) spectroscopy in approximately 121,000 UK Biobank participants has recently been released as a community resource, comprising absolute concentrations and ratios of 249 circulating metabolites, lipids, and lipoprotein sub-fractions. Here we identify and characterise additional sources of unwanted technical variation influencing individual biomarkers in the data available to download from UK Biobank. These included sample preparation time, shipping plate well, spectrometer batch effects, drift over time within spectrometer, and outlier shipping plates. We developed a procedure for removing this unwanted technical variation, and demonstrate that it increases signal for genetic and epidemiological studies of the NMR metabolic biomarker data in UK Biobank. We subsequently developed an R package, ukbnmr, which we make available to the wider research community to enhance the utility of the UK Biobank NMR metabolic biomarker data and to facilitate rapid analysis.

Cross-sectionally calculated metabolic ageing does not relate to longitudinal metabolic changes - support for stratified ageing models

CONTEXT

Ageing varies between individuals with profound consequences for chronic diseases and longevity. One hypothesis to explain the diversity is a genetically regulated molecular clock that runs differently between individuals. Large and long enough human studies to test the hypothesis are rare due to practical challenges, but statistical models of ageing are built as proxies for the molecular clock by comparing young and old individuals cross-sectionally. These models remain untested against longitudinal data.

OBJECTIVE

We applied novel methodology to test if cross-sectional modelling can distinguish slow versus accelerated ageing in a human population.

DESIGN

We trained a machine learning model to predict age from 153 clinical and cardiometabolic traits. The model was tested against longitudinal data from another cohort.

PATIENTS OR OTHER PARTICIPANTS

The training data came from cross-sectional surveys of the Finnish population (n = 9,708; ages 25-74 years). The validation data included three time points across 10 years in the Young Finns Study (YFS; n = 1,009; ages 24-49 years).

INTERVENTION(S)

Predicted metabolic age in 2007 was compared against observed ageing rate from the 2001 visit to the 2011 visit in the YFS dataset.

MAIN OUTCOME MEASURE(S)

Correlation between predicted versus observed metabolic ageing.

RESULTS

The cross-sectional proxy failed to predict longitudinal observations (R2 = 0.018%, P = 0.67).

CONCLUSIONS

The finding is unexpected under the clock hypothesis that would produce a positive correlation between predicted and observed ageing. Our results are better explained by a stratified model where ageing rates per se are similar in adulthood but differences in starting points explain diverging metabolic fates.

Association of Advanced Lipoprotein Subpopulation Profiles with Insulin Resistance and Inflammation in Patients with Type 2 Diabetes Mellitus

Plasma lipoproteins exist as several subpopulations with distinct particle number and size that are not fully reflected in the conventional lipid panel. In this study, we sought to quantify lipoprotein subpopulations in patients with type 2 diabetes mellitus (T2DM) to determine whether specific lipoprotein subpopulations are associated with insulin resistance and inflammation markers. The study included 57 patients with T2DM (age, 61.14 ± 9.99 years; HbA1c, 8.66 ± 1.60%; mean body mass index, 35.15 ± 6.65 kg/m²). Plasma lipoprotein particles number and size were determined by nuclear magnetic resonance spectroscopy. Associations of different lipoprotein subpopulations with lipoprotein insulin resistance (LPIR) score and glycoprotein acetylation (GlycA) were assessed using multi-regression analysis. In stepwise regression analysis, VLDL and HDL large particle number and size showed the strongest associations with LPIR (R² = 0.960; p = 0.0001), whereas the concentrations of the small VLDL and HDL particles were associated with GlycA (R² = 0.190; p = 0.008 and p = 0.049, respectively). In adjusted multi-regression analysis, small and large VLDL particles and all sizes of lipoproteins independently predicted LPIR, whereas only the number of small LDL particles predicted GlycA. Conventional markers HbA1c and Hs-CRP did not exhibit any significant association with lipoprotein subpopulations. Our data suggest that monitoring insulin resistance-induced changes in lipoprotein subpopulations in T2DM might help to identify novel biomarkers that can be useful for effective clinical intervention.

High-density lipoprotein subclasses and cardiovascular disease and mortality in type 2 diabetes: analysis from the Hong Kong Diabetes Biobank

OBJECTIVE

High-density lipoproteins (HDL) comprise particles of different size, density and composition and their vasoprotective functions may differ. Diabetes modifies the composition and function of HDL. We assessed associations of HDL size-based subclasses with incident cardiovascular disease (CVD) and mortality and their prognostic utility.

RESEARCH DESIGN AND METHODS

HDL subclasses by nuclear magnetic resonance spectroscopy were determined in sera from 1991 fasted adults with type 2 diabetes (T2D) consecutively recruited from March 2014 to February 2015 in Hong Kong. HDL was divided into small, medium, large and very large subclasses. Associations (per SD increment) with outcomes were evaluated using multivariate Cox proportional hazards models. C-statistic, integrated discrimination index (IDI), and categorial and continuous net reclassification improvement (NRI) were used to assess predictive value.

RESULTS

Over median (IQR) 5.2 (5.0-5.4) years, 125 participants developed incident CVD and 90 participants died. Small HDL particles (HDL-P) were inversely associated with incident CVD [hazard ratio (HR) 0.65 (95% CI 0.52, 0.81)] and all-cause mortality [0.47 (0.38, 0.59)] (false discovery rate < 0.05). Very large HDL-P were positively associated with all-cause mortality [1.75 (1.19, 2.58)]. Small HDL-P improved prediction of mortality [C-statistic 0.034 (0.013, 0.055), IDI 0.052 (0.014, 0.103), categorical NRI 0.156 (0.006, 0.252), and continuous NRI 0.571 (0.246, 0.851)] and CVD [IDI 0.017 (0.003, 0.038) and continuous NRI 0.282 (0.088, 0.486)] over the RECODe model.

CONCLUSION

Small HDL-P were inversely associated with incident CVD and all-cause mortality and improved risk stratification for adverse outcomes in people with T2D. HDL-P may be used as markers for residual risk in people with T2D.

Longitudinal profiling of metabolic ageing trends in two population cohorts of young adults

BACKGROUND

Quantification of metabolic changes over the human life course is essential to understanding ageing processes. Yet longitudinal metabolomics data are rare and long gaps between visits can introduce biases that mask true trends. We introduce new ways to process quantitative time-series population data and elucidate metabolic ageing trends in two large cohorts.

METHODS

Eligible participants included 1672 individuals from the Cardiovascular Risk in Young Finns Study and 3117 from the Northern Finland Birth Cohort 1966. Up to three time points (ages 24-49 years) were analysed by nuclear magnetic resonance metabolomics and clinical biochemistry (236 measures). Temporal trends were quantified as median change per decade. Sample quality was verified by consistency of shared biomarkers between metabolomics and clinical assays. Batch effects between visits were mitigated by a new algorithm introduced in this report. The results below satisfy multiple testing threshold of P < 0.0006.

RESULTS

Women gained more weight than men (+6.5% vs +5.0%) but showed milder metabolic changes overall. Temporal sex differences were observed for C-reactive protein (women +5.1%, men +21.1%), glycine (women +5.2%, men +1.9%) and phenylalanine (women +0.6%, men +3.5%). In 566 individuals with ≥+3% weight gain vs 561 with weight change ≤-3%, divergent patterns were observed for insulin (+24% vs -10%), very-low-density-lipoprotein triglycerides (+32% vs -6%), high-density-lipoprotein2 cholesterol (-6.5% vs +4.7%), isoleucine (+5.7% vs -6.0%) and C-reactive protein (+25% vs -22%).

CONCLUSION

We report absolute and proportional trends for 236 metabolic measures as new reference material for overall age-associated and specific weight-driven changes in real-world populations.

Metabolomic profiles predict individual multidisease outcomes

Risk stratification is critical for the early identification of high-risk individuals and disease prevention. Here we explored the potential of nuclear magnetic resonance (NMR) spectroscopy-derived metabolomic profiles to inform on multidisease risk beyond conventional clinical predictors for the onset of 24 common conditions, including metabolic, vascular, respiratory, musculoskeletal and neurological diseases and cancers. Specifically, we trained a neural network to learn disease-specific metabolomic states from 168 circulating metabolic markers measured in 117,981 participants with ~1.4 million person-years of follow-up from the UK Biobank and validated the model in four independent cohorts. We found metabolomic states to be associated with incident event rates in all the investigated conditions, except breast cancer. For 10-year outcome prediction for 15 endpoints, with and without established metabolic contribution, a combination of age and sex and the metabolomic state equaled or outperformed established predictors. Moreover, metabolomic state added predictive information over comprehensive clinical variables for eight common diseases, including type 2 diabetes, dementia and heart failure. Decision curve analyses showed that predictive improvements translated into clinical utility for a wide range of potential decision thresholds. Taken together, our study demonstrates both the potential and limitations of NMR-derived metabolomic profiles as a multidisease assay to inform on the risk of many common diseases simultaneously.

Gut Microbiome and Its Cofactors Are Linked to Lipoprotein Distribution Profiles

Increasing evidence indicates that the gut microbiome (GM) plays an important role in dyslipidemia. To date, however, no in-depth characterization of the associations between GM with lipoproteins distributions (LPD) among adult individuals with diverse BMI has been conducted. To determine such associations, we studied blood-plasma LPD, fecal short-chain fatty acids (SCFA) and GM of 262 Danes aged 19–89 years. Stratification of LPD segregated subjects into three clusters displaying recommended levels of lipoproteins and explained by age and body-mass-index. Higher levels of HDL2a and HDL2b were associated with a higher abundance of Ruminococcaceae and Christensenellaceae. Increasing levels of total cholesterol and LDL-1 and LDL-2 were positively associated with Lachnospiraceae and Coriobacteriaceae, and negatively with Bacteroidaceae and Bifidobacteriaceae. Metagenome-sequencing showed a higher abundance of biosynthesis of multiple B-vitamins and SCFA metabolism genes among healthier LPD profiles. Metagenomic-assembled genomes (MAGs) affiliated to Eggerthellaceae and Clostridiales were contributors of these genes and their relative abundance correlated positively with larger HDL subfractions. The study demonstrates that differences in composition and metabolic traits of the GM are associated with variations in LPD among the recruited subjects. These findings provide evidence for GM considerations in future research aiming to shed light on mechanisms of the GM–dyslipidemia axis.

Constructing an atlas of associations between polygenic scores from across the human phenome and circulating metabolic biomarkers

Background

Polygenic scores (PGS) are becoming an increasingly popular approach to predict complex disease risk, although they also hold the potential to develop insight into the molecular profiles of patients with an elevated genetic predisposition to disease.

Methods

We sought to construct an atlas of associations between 125 different PGS derived using results from genome-wide association studies and 249 circulating metabolites in up to 83,004 participants from the UK Biobank.

Results

As an exemplar to demonstrate the value of this atlas, we conducted a hypothesis-free evaluation of all associations with glycoprotein acetyls (GlycA), an inflammatory biomarker. Using bidirectional Mendelian randomization, we find that the associations highlighted likely reflect the effect of risk factors, such as adiposity or liability towards smoking, on systemic inflammation as opposed to the converse direction. Moreover, we repeated all analyses in our atlas within age strata to investigate potential sources of collider bias, such as medication usage. This was exemplified by comparing associations between lipoprotein lipid profiles and the coronary artery disease PGS in the youngest and oldest age strata, which had differing proportions of individuals undergoing statin therapy. Lastly, we generated all PGS-metabolite associations stratified by sex and separately after excluding 13 established lipid-associated loci to further evaluate the robustness of findings.

Conclusions

We envisage that the atlas of results constructed in our study will motivate future hypothesis generation and help prioritize and deprioritize circulating metabolic traits for in-depth investigations. All results can be visualized and downloaded at http://mrcieu.mrsoftware.org/metabolites_PRS_atlas.

Funding

This work is supported by funding from the Wellcome Trust, the British Heart Foundation, and the Medical Research Council Integrative Epidemiology Unit.

Constructing an atlas of associations between polygenic scores from across the human phenome and circulating metabolic biomarkers

BACKGROUND

Polygenic scores (PGS) are becoming an increasingly popular approach to predict complex disease risk, although they also hold the potential to develop insight into the molecular profiles of patients with an elevated genetic predisposition to disease.

METHODS

We sought to construct an atlas of associations between 125 different PGS derived using results from genome-wide association studies and 249 circulating metabolites in up to 83,004 participants from the UK Biobank.

RESULTS

As an exemplar to demonstrate the value of this atlas, we conducted a hypothesis-free evaluation of all associations with glycoprotein acetyls (GlycA), an inflammatory biomarker. Using bidirectional Mendelian randomization, we find that the associations highlighted likely reflect the effect of risk factors, such as adiposity or liability towards smoking, on systemic inflammation as opposed to the converse direction. Moreover, we repeated all analyses in our atlas within age strata to investigate potential sources of collider bias, such as medication usage. This was exemplified by comparing associations between lipoprotein lipid profiles and the coronary artery disease PGS in the youngest and oldest age strata, which had differing proportions of individuals undergoing statin therapy. Lastly, we generated all PGS-metabolite associations stratified by sex and separately after excluding 13 established lipid-associated loci to further evaluate the robustness of findings.

CONCLUSIONS

We envisage that the atlas of results constructed in our study will motivate future hypothesis generation and help prioritize and deprioritize circulating metabolic traits for in-depth investigations. All results can be visualized and downloaded at http://mrcieu.mrsoftware.org/metabolites_PRS_atlas.

FUNDING

This work is supported by funding from the Wellcome Trust, the British Heart Foundation, and the Medical Research Council Integrative Epidemiology Unit.

Physical activity and health: Findings from Finnish monozygotic twin pairs discordant for physical activity

Genetic and early environmental differences including early health habits associate with future health. To provide insight on the causal nature of these associations, monozygotic (MZ) twin pairs discordant for health habits provide an interesting natural experiment. Twin pairs discordant for leisure‐time physical activity (LTPA) in early adult life is thus a powerful study design to investigate the associations between long‐term LTPA and indicators of health and wellbeing. We have identified 17 LTPA discordant twin pairs from two Finnish twin cohorts and summarize key findings of these studies in this paper. The carefully characterized rare long‐term LTPA discordant MZ twin pairs have participated in multi‐dimensional clinical examinations. Key findings highlight that compared with less active twins in such MZ twin pairs, the twins with higher long‐term LTPA have higher physical fitness, reduced body fat, reduced visceral fat, reduced liver fat, increased lumen diameters of conduit arteries to the lower limbs, increased bone mineral density in loaded bone areas, and an increased number of large high‐density lipoprotein particles. The findings increase our understanding on the possible site‐specific and system‐level effects of long‐term LTPA.