Very Low Density Lipoprotein

A high-throughput serum nuclear magnetic resonance (NMR) spectroscopy platform [28] was utilized to quantify 67 metabolic measures that represent a broad molecular signature of the systemic metabolite profile. The metabolite set covers multiple metabolic pathways, and includes lipoprotein lipids, fatty acids, amino acids, and glycolysis precursors (Table S1). Fourteen lipoprotein subclasses were analyzed as part of the metabolite profile, with the subclass sizes defined as follows: extremely large very-low-density lipoproteins (VLDLs) (particle diameter from 75 nm upwards), five VLDL subclasses (average particle diameters of 64.0 nm, 53.6 nm, 44.5 nm, 36.8 nm, and 31.3 nm), intermediate-density lipoproteins (28.6 nm), three low-density lipoprotein (LDL) subclasses (25.5 nm, 23.0 nm, and 18.7 nm), and four high-density lipoprotein (HDL) subclasses (14.3 nm, 12.1 nm, 10.9 nm, and 8.7 nm). The NMR-based metabolite profiling employed in this study has previously been used in various epidemiological studies [25] (link)–[31] (link), and details of the experimentation have been described [28] ,[32] (link),[33] (link). Furthermore, 15 additional measures, including various inflammatory markers, liver function surrogates, hormones, and blood pressure, were analyzed (Text S1). These additional metabolic measures, assayed in at least two of the cohorts, were selected to complement the comprehensive characterization of cardiometabolic effects of adiposity across multiple pathways and to enhance comparability with prior Mendelian randomization studies [7] (link),[14] (link)–[16] (link).

Blood samples were collected from the antecubital vein of all participants in the morning under fasting conditions. They were stored in vacuum tubes containing EDTA (ethylene diamine tetraacetic acid) and coagulation tubes. A range of haematological and biochemistry tests (Table 2) were conducted on fresh samples at the central laboratory of the Staff Hospital of Jidong oil-field of Chinese National Petroleum. Fasting blood glucose was measured with the hexokinase/glucose-6-phosphate dehydrogenase method. Cholesterol and triglyceride concentrations were determined by enzymatic methods (Mind Bioengineering Co. Ltd, Shanghai, China). Blood samples were also measured using an auto-analyzer (Hitachi 747; Hitachi, Tokyo, Japan) at the central laboratory of the Staff Hospital of Jidong oil-field of Chinese National Petroleum. For all participants, serum creatinine, cholesterol, high-density lipoproteins (HDL-C), low-density lipoproteins (LDL-C), triglycerides and glucose levels were assessed. In subgroup analysis studies, various biomarkers of blood cells, serum and plasma were measured: C-reactive protein, homocysteine, estrogens, androgens, vitamin D, lipoprotein-associated phospholipase A2 (Lp-PLA2), insulin, and glycosylated hemoglobin HbA1c.Table 2

Haematology, biochemistry and biological specimen banking in the COACS

	Analysate
Red blood cells	Haemoglobin
	Red corpuscle count
	Haematocrit
	Mean corpuscular volume
	Mean corpuscular
	Haemoglobin concentration
	Red blood cell distribution width
White blood cells	White cell count Total count
	Differential count
Platelets	Platelets Count
	Mean platelet volume
Urea	Urine specific gravity
	Ery
	Urea nitrogen
	Uric acid (UA)
	Creatinine (Cr)
	Urine protein
Liver function tests (plasma)	Alkaline phosphatise
	Alanine transaminase (ALT)
	Aspartate aminotransferase (AST)
	Phosphatise Transglutaminase (TG)
Liver function tests (serum)	HBsAg
	Anti-HBs
	HBeAg
	Anti-HBe
	Anti-HBc
Lipids (plasma)	Total cholesterol (TC)
	Total bilirubin (TBIL)
	Triglycerides (TG)
	Low density lipoprotein (LDL)
	Very Low density lipoprotein (VLDL)
General chemistry (plasma)	C-reactive protein
	Homocysteine
	Steroids
	Glucose
	Insulin
	Glycosylated hemoglobin
Bio-specimen banking
White blood cells	DNA, RNA extraction and analyses
Serum	Pedtidome profiling
Plasma	Glycome

Blood samples were processed and separated onsite for biospecimen banking (−80 °C). DNA and RNA were extracted and stored in the laboratory of Beijing Key Laboratory of Clinical Epidemiology, Beijing, China.

Body composition and blood pressure were assessed by a research nurse at Institute of Biomedicine, University of Eastern Finland. Body height was measured by a wall-mounted stadiometer and body weight by the Inbody 720® bioimpedance device
[23 (link)]. Body mass index (BMI) was calculated as body weight divided by body height squared. BMI-standard deviation score (BMI-SDS) was assessed by national references
[24 (link)]. Body fat percentage and lean mass were measured by the Lunar® dual energy x-ray absorptiometry (DXA) device
[23 (link)]. Waist circumference was measured after expiration at mid-distance between the bottom of the rib cage and the top of the iliac crest
[23 (link)]. Blood pressure was measured manually on the right arm by a calibrated aneroid sphygmomanometer (Heine 130 Gamma G7, Munich, Germany). The measurement protocol included, after a rest of five minutes, three measurements in the sitting position at 2-minute intervals. The mean of all three values were used as the systolic and diastolic blood pressure.
Venous blood samples were taken after a 12-hour overnight fast by a laboratory nurse at Institute of Biomedicine, University of Eastern Finland. The blood samples were analysed at Laboratory of Clinical Chemistry at University Hospital of Kuopio. The assessment of plasma glucose, total, high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol and triglycerides as well as serum insulin from 12-hour fasting samples has been explained previously
[25 (link)]. Very low-density lipoprotein (VLDL) was separated by ultracentrifugation and HDL by precipitation of LDL after removal of VLDL fraction
[26 (link)].
A continuous cardiometabolic risk score variable was calculated as the sum of Z-scores of waist circumference, insulin, glucose, triglycerides, HDL cholesterol and the mean of systolic and diastolic blood pressure that are specific for the PANIC study population. The Z-score of HDL cholesterol was multiplied by -1, because HDL cholesterol is inversely associated with cardiometabolic risk. A higher cardiometabolic risk score indicates a less favourable cardiometabolic risk profile.

Mice at the fed state or fasted for 12 h were sacrificed, and whole blood was collected from retrobulbar venous plexus and centrifuged for 10 min at 12,000 × g to obtain plasma. Plasma glucose, total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), very low-density lipoprotein (VLDL), and total triglycerides were measured using commercial kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). ELISA kits (Sinoukbio, Beijing, China) were used to measure plasma insulin, C-peptide, non-esterified fatty acid (NEFA), leptin, IL (interleukin) 4, IL6, IL10, resistin, interferon γ (IFNγ) and monocyte chemotactic protein-1 (MCP1) following the manufacturer’s instruction.

The levels of total cholesterol (TC), triglycerides (TG), and high-density lipoprotein (HDL) were carried out in semi autoanalyzer (Photometer 5010, Germany) using Agappe Kits. Friedwalds formula was used to calculate the levels of low-density lipoprotein (LDL) and very low-density lipoprotein (VLDL) as follows:

All subjects underwent overnight sleep monitoring. In the study, only patients with polysomnography (Alice 6 device, Philips-Respironics, Murrysville, PA, USA) and recorded respiratory disturbance index (RDI) were enrolled. Only subjects with RDI ≥ 15 were included for further assessment. Other recorded indices included oxygen desaturation index (ODI), arousal index, average nocturnal O₂ saturation, and minimal nocturnal O₂ saturation. Standardized criteria were used for the scoring of sleep characteristics and respiratory events [22 ].
Monitored variables:
Blood plasma samples were obtained in the morning after polysomnography and after overnight fasting. Blood samples with ethylenediaminetetraacetic acid (EDTA) were collected. Immediately after the collection of plasma samples, levels of TAG, TC, LDL, and HDL were determined in a local certified hospital laboratory with an enzymatic method (Roche Diagnostics, Mannheim, Germany). The quantitative analysis of lipoprotein families and lipoprotein subfractions including very low-density lipoprotein (VLDL), intermediate-density lipoprotein (IDL), and plasma lipoprotein subfractions were analyzed by the Lipoprotein system (Quantimetrix Corp., Redondo Beach, CA, USA) using a polyacrylamide gel electrophoresis [23 (link)]. The following subfractions were evaluated: large LDL subfractions 1–2 (which are considered atheroprotective), small dense LDL subfractions 3–7 (which are considered atherogenic), large HDL subfractions 1–3 (which are considered atheroprotective), small dense HDL subfractions 8–10 (which are considered atherogenic), and intermediate HDL subfractions 4–7 (their atherogenic/atheroprotective role remains controversial) [20 (link),21 (link)].
Endothelial function was assessed by PAT (EndoPAT 2000 device, Itamar Medical Ltd., Caesarea, Israel) as previously described [14 (link)]. RHI was calculated as the ratio of the average amplitude of the PAT signal post-to-pre occlusion of the tested arm, normalized to the concurrent signal from the contralateral finger. Calculations were performed using the computer algorithm (software 3.1.2) supplied with the device. RHI value < 1.67 indicated endothelial dysfunction [14 (link),24 (link)].
Statistical analyses were performed by SPSS ver. 18 (SPSS Inc., Chicago, IL, USA). The results of normally distributed data are expressed as a mean ± standard deviation, and the results of not normally distributed data are expressed as median, interquartile range, minimal and maximal values. Pearson or Spearman correlation coefficients were used to determine the relationships between RHI and the baseline characteristics of the study population. We used stepwise multiple linear regression to create the prediction model and identify the most important contributors to this model. A model with the highest number of significant predictors was chosen. The dependent variable in the model was RHI, independent variables in the model were anthropometric characteristics (age, gender, BMI), sleep characteristics (T90, RDI, ODI, arousal index, average, and minimal nocturnal O₂ saturation), and lipoprotein levels (TAG, TC, LDL, HDL, VLDL, IDL, large LDL, small LDL, large HDL, intermediate HDL, and small HDL). Each model was assessed for the presence of multicollinearity of included variables. The variance inflation factor (VIF) ≥5 was indicative of multicollinearity. The p value < 0.05 was considered statistically significant.