| Literature DB >> 34902008 |
Yuanjie Pang1, Christiana Kartsonaki2,3, Jun Lv1,4, Iona Y Millwood2,3, Zammy Fairhurst-Hunter2, Iain Turnbull2, Fiona Bragg2,3, Michael R Hill2, Canqing Yu1,4, Yu Guo5, Yiping Chen2,3, Ling Yang2,3, Robert Clarke2, Robin G Walters2,3, Ming Wu6, Junshi Chen7, Liming Li1,4, Zhengming Chen2,3, Michael V Holmes2,3,8.
Abstract
BACKGROUND: Globally, the burden of obesity and associated nonalcoholic fatty liver disease (NAFLD) are rising, but little is known about the role that circulating metabolomic biomarkers play in mediating their association.Entities:
Keywords: Chinese; Mendelian randomization; adiposity; metabolomics; nonalcoholic fatty liver disease
Mesh:
Substances:
Year: 2022 PMID: 34902008 PMCID: PMC8895224 DOI: 10.1093/ajcn/nqab392
Source DB: PubMed Journal: Am J Clin Nutr ISSN: 0002-9165 Impact factor: 8.472
FIGURE 1Flow diagram of the study design. CLD, chronic liver disease; CVD, cardiovascular disease; GWAS, genome-wide association study; MR, Mendelian randomization; NAFLD, nonalcoholic fatty liver disease.
Baseline characteristics of subcohort participants and NAFLD cases[1]
| Subcohort | NAFLD cases | |
|---|---|---|
| Variable | ( | ( |
| Age (±SD), y | 50.6 ± 10.1 | 51.1 ± 9.1 |
| Female, % | 62.8 | 64.5 |
| Socioeconomic and lifestyle factors, % | ||
| Urban residency | 48.6 | 25.1 |
| ≥9 years of education | 21.6 | 28.0 |
| Household income ≥35 000 RMB/y | 19.5 | 21.4 |
| Ever regular smoking, % | ||
| Male | 65.1 | 69.1 |
| Female | 5.9 | 9.2 |
| Weekly alcohol drinking, % | ||
| Male | 31.6 | 45.2 |
| Female | 4.5 | 4.6 |
| Total physical activity (±SD), MET-h/d | 21.8 ± 14.6 | 21.2 ± 13.4 |
| Sedentary leisure time (±SD), h/d | 3.1 ± 1.5 | 3.0 ± 1.9 |
| Blood pressure and anthropometry | ||
| SBP (±SD), mmHg | 131.2 ± 19.9 | 133.0 ± 19.7 |
| RPG (±SD), mmol/L | 5.9 ± 2.7 | 6.4 ± 2.5 |
| BMI (±SD), kg/m2 | 23.9 ± 3.5 | 26.1 ± 3.5 |
| Waist circumference (±SD), cm | 79.6 ± 10.0 | 85.7 ± 9.9 |
| Hip circumference (±SD), cm | 90.5 ± 7.3 | 94.0 ± 7.2 |
| Waist-to-hip ratio (±SD) | 0.88 ± 0.07 | 0.91 ± 0.08 |
| Percentage body fat (±SD), % | 29.0 ± 8.2 | 33.4 ± 9.2 |
| Standing height (±SD), cm | 158.2 ± 7.9 | 158.7 ± 9.1 |
| Prior disease history, % | ||
| Diabetes | 5.8 | 6.5 |
| Coronary heart disease | 3.1 | 5.9 |
| Stroke or TIA | 2.9 | 0 |
| Hypertension | 10.9 | 13.7 |
| Family history of diabetes | 5.3 | 5.3 |
| Family history of cancer | 18.5 | 18.6 |
Results by BMI categories are standardized by age, sex, and region, whereas for age they were adjusted for sex and region. Values are means unless otherwise stated. MET, metabolic equivalent of task; NAFLD, nonalcoholic fatty liver disease; RMB, Renminbi; RPG, random plasma glucose; SBP, systolic blood pressure; TIA, transient ischemic attack.
FIGURE 2Manhattan plot showing the P values for observational associations of adiposity, metabolomic biomarkers, and risk of NAFLD. Observational associations of BMI and WC with all metabolomic biomarkers (A, B) and of metabolomic biomarkers with NAFLD risk (C) are shown. Metabolomic biomarkers are the dependent variables in A and B, whereas NAFLD is the dependent variable in C. The x axis shows the numeric order of metabolomic biomarkers shown in “Compound ID” Supplemental Table 1 within each super-pathway. The y axis shows log-transformed FDR-corrected P values. The horizontal line denotes −log(0.05). The table shows the percentage of metabolomic biomarkers that passed the FDR threshold of 0.05 by super-pathways. The estimates were adjusted for age, age squared, sex, area, smoking, education, and fasting time. FDR, false discovery rate; NAFLD, nonalcoholic fatty liver disease; WC, waist circumference. An asterisk denotes that the metabolic biomarker is a mixture of isoforms (with the same number of carbons and carbon-carbon double bonds).
FIGURE 3Associations of BMI, metabolomic biomarkers, and risk of NAFLD for 15 metabolomic biomarkers associated with both BMI and NAFLD at 5% FDR. Column A shows adjusted SD differences (95% CI) of metabolomic biomarkers per 1-SD higher observational BMI for 15 metabolomic biomarkers associated with both BMI and NAFLD with FDR-corrected P values <0.05. Column B shows corresponding estimates per 1-SD higher genetically elevated BMI. The observational estimates were adjusted for age, age squared, sex, area, smoking, education, and fasting time. The MR estimates were adjusted for age, age squared, sex, area, the first 12 principal components, education, and smoking. The SD for BMI in the whole CKB cohort was 3.4 kg/m2. Column C shows adjusted HR (95% CI) of NAFLD per 1-SD higher metabolomic biomarkers. An asterisk denotes that the metabolic biomarker is a mixture of isoforms (with the same number of carbons and carbon-carbon double bonds). FDR, false discovery rate; GPG, glycerophosphoglycerol; GPC, glycerophosphocholine; OB, observational; MR, Mendelian randomization; NAFLD, nonalcoholic fatty liver disease.
Exploratory investigation of metabolomic biomarkers as predictors of incident diagnosis of NAFLD[1]
| Variables included in model | Weighted C statistic | 95% CI |
|
|---|---|---|---|
| Base model | 0.84 | (0.80, 0.88) | |
| + 15 metabolic biomarkers (criterion 1) | 0.88 | (0.85, 0.92) | 0.005 |
| + 21 metabolic biomarkers (criteria 1, 2) | 0.89 | (0.86, 0.93) | 0.002 |
| + 28 metabolic biomarkers (criteria 1–3) | 0.90 | (0.87, 0.93) | <0.001 |
1Base model includes age, age squared, sex, region, education, household income, smoking, total physical activity, BMI, and diabetes. Discrimination of models was assessed using a weighted C-index.
The criterion was each defined as 1) those with z >2.5, 2) the 2 most significant, or 3) the 3 most significant. P values are for comparison with the base model. NAFLD, nonalcoholic fatty liver disease.
FIGURE 4Central illustration of BMI, metabolomic biomarkers, and risk of NAFLD. We assessed the associations of BMI with ∼1200 metabolomic biomarkers and of these metabolomic biomarkers with risk of NAFLD. Previous reports in CKB showed observational and genetic associations between BMI and risk of NAFLD (21, 62). For the associations between BMI and metabolomic biomarkers, this study showed that measured BMI was observationally associated with 199 metabolomic biomarkers (amino acids, carbohydrates, cofactors and vitamins, lipids, nucleotides, peptides, and xenobiotics), with general concordance between the observational and genetic associations (except for 3 biomarkers—i.e., 1,5-anhydroglucitol, N-acetylglycine, and C10H18O2). This study also showed that 35 metabolomic biomarkers were associated with NAFLD risk. The lower panel illustrates the observational associations of BMI with metabolomics and of metabolomics with NAFLD by sub-pathways. There were 15 metabolomic biomarkers that were associated with BMI and NAFLD risk. CKB, China Kadoorie Biobank; HR, hazard ratio; NAFLD, nonalcoholic fatty liver disease.