Effects of sleep restriction on glucose control and insulin secretion during diet-induced weight loss

Insufficient sleep is associated with changes in glucose tolerance, insulin secretion, and insulin action. Despite widespread use of weight-loss diets for metabolic risk reduction, the effects of insufficient sleep on glucose regulation in overweight dieters are not known. To examine the consequences of recurrent sleep restriction on 24-h blood glucose control during diet-induced weight loss, 10 overweight and obese adults (3F/7M; mean (s.d.) age 41 (5) years; BMI 27.4 (2.0) kg/m(2)) completed two 14-day treatments with hypocaloric diet and 8.5- or 5.5-h nighttime sleep opportunity in random order 7 (3) months apart. Oral and intravenous glucose tolerance test (IVGTT) data, fasting lipids and free fatty acids (FFA), 24-h blood glucose, insulin, C-peptide, and counter-regulatory hormone measurements were collected after each treatment. Participants had comparable weight loss (1.0 (0.3) BMI units) during each treatment. Bedtime restriction reduced sleep by 131 (30) min/day. Recurrent sleep curtailment decreased 24-h serum insulin concentrations (i.e., enhanced 24-h insulin economy) without changes in oral glucose tolerance and 24-h glucose control. This was accompanied by a decline in fasting blood glucose, increased fasting FFA, which suppressed normally following glucose ingestion, and lower total and low-density lipoprotein cholesterol concentrations. Sleep-loss-related changes in counter-regulatory hormone secretion during the IVGTT limited the utility of the test in this study. In conclusion, sleep restriction enhanced 24-h insulin economy without compromising glucose homeostasis in overweight individuals placed on a balanced hypocaloric diet. The changes in fasting blood glucose, insulin, lipid and FFA concentrations in sleep-restricted dieters resembled the pattern of human metabolic adaptation to reduced carbohydrate availability.     

Effect of propranolol on glucose disposal rate in patients with hyperthyroidism

The effect of propranolol treatment (60 mg per day, three days) on glucose disposal rate (K-value) was investigated in nine patients with hyperthyroidism. K-value was improved in 4 cases and aggravated in 5 cases. The fasting levels of plasma free fatty acids (FFA) before the administration of propranolol in the improved cases were significantly higher than those in the aggravated cases. The propranolol treatment markedly reduced FFA levels only in the improved cases. These results suggest that the impaired glucose tolerance frequently seen in hyperthyroidism patients could be partly attributed to the increased level of plasma FFA.     

Individual plasma ghrelin changes in the same patients in hyperthyroid,hypothyroid and euthyroid state

Ghrelin is a multifunctional peptide of widespread expression. Since it has been shown to influence energy homeostatis, its potential role in thyroid dysfunction may have clinical significance. In this study, plasma ghrelin changes have been analyzed in the same patients in three different thyroid states for the first time. The study group consisted of 16 patients who had been diagnosed with hyperthyroidism, were treated with radioiodine, developed hypothyroidism after treatment, and finally became euthyroid on l-thyroxine substitution. In the initial state of hyperthyroidism plasma ghrelin levels correlated negatively with fT₃ and fT₄. In hypothyroidism ghrelin concentration increased significantly (p < 0.05). Although the mean value of plasma ghrelin tended to decrease in the euthyroid state, the individual difference between hypothyroidism and euthyroidism was not significant. Plasma ghrelin in euthyroidism was still significantly higher than in hyperthyroidism (p < 0.05), and correlated positively with ghrelin levels in hyperthyroidism and hypothyroidism. In our opinion, plasma ghrelin fluctuations may reflect metabolic changes in patients with thyroid dysfunction. Moreover, it cannot be excluded that in thyroid disorders ghrelin acts as a compensatory factor, helping to balance metabolic disturbances.     

Factors predicting the course of diabetes mellitus in hyperthyroid patients

The relationship between changes in glucose tolerance with treatment of hyperthyroidism and various factors that might be relevant to carbohydrate metabolism were investigated in 64 hyperthyroid patients with abnormal glucose tolerance, including 35 cases with fasting plasma glucose (FPG) levels of 140 mg/dl or more. All patients had diffuse toxic goiter. After correction of the hyperthyroidism, glucose intolerance improved in almost all cases, even in cases with fasting hyperglycemia, but diabetes mellitus in patients with FPG above 140 mg/dl and/or delta IRI/delta PG X 30' during a 50-g oral glucose tolerance test below 0.10, persisted. Patients who showed diabetic glucose tolerance even after remission from thyroid dysfunction had significantly lower delta IRI/delta PG X 30' values and a higher incidence of family histories of diabetes mellitus than those not showing diabetic glucose tolerance. There were no significant differences in serum T3 and T4 levels between these two groups of patients. The findings suggest that predisposition to diabetes may be an important factor in persistent glucose intolerance in the hyperthyroidism of Graves' disease. The FPG and delta IRI/delta PG X 30' values may be useful in predicting which patients with hyperthyroidism will have permanent diabetes.     

Screening of geriatric patients for thyroid dysfunction with thyrotropin-releasing-hormone test, sensitive thyrotropin and free thyroxine estimation

Hospitalized geriatric patients (N = 354) from an iodine-deficient area were screened with sensitive thyrotropin (TSH), free and total thyroxine (FT4, T4) and total triiodothyronine (T3) to determine the occurrence rate of clinical and subclinical thyroid dysfunction. The diagnostic value of the tests was compared to each other and to that of the thyrotropin-releasing-hormone test (TRH-test) in order to find the optimal first line screening test in geriatric patients. Clinical hyperthyroidism was found in 13, subclinical hyperthyroidism in 10, overt hypothyroidism in 6 and subclinical hypothyroidism in 8 cases. 20.6% of the patients were euthyroid but had subnormal TSH response to TRH, as a sign of possible thyroid autonomy. The low occurrence rate of clinical thyroid disorders (4.8%) does not justify the screening of geriatric patients in general, but the high probability of thyroid autonomy makes reasonable the investigation of every geriatric patient before iodine administration. Suppressed basal TSH and high FT4 were found to be both sensitive and specific in diagnosing clinical hyperthyroidism, but the predictive value was insufficient; elevated T4 and T3 are specific, but not sensitive. Basal TSH is sensitive, specific and has a good predictive value in diagnosing euthyroidism, whereas normal T4, FT4 or T3 are not specific enough for euthyroidism. Basal TSH is better as a first line test of thyroid function than FT4. A normal basal TSH confirms euthyroidism by itself. Other tests (TRH test, T4, FT4, T3) are necessary to elucidate the clinical importance of a subnormal or suppressed basal TSH.     

Overwinter fasting and re-feeding in rainbow trout: plasma growth hormone and cortisol levels in relation to energy mobilisation

This study investigated the roles of cortisol and growth hormone (GH) during a period of fasting in overwintering salmonid fish. Indices of carbohydrate (plasma glucose, liver glycogen), lipid (plasma free fatty acids (FFAs)) and protein metabolism (plasma protein, total plasma amino acids) were determined, together with plasma GH, cortisol and somatolactin (SL) levels at intervals in three groups of rainbow trout (continuously fed; fasted for 9 weeks then fed; fasted for 17 weeks). In fasted fish, a decline in body weight and condition factor was accompanied by reduced plasma glucose and hepatic glycogen and increased plasma FFA. No consistent elevation of plasma GH occurred until after 8 weeks of fasting when plasma GH levels increased ninefold. No changes were observed in plasma total protein and AA until between weeks 13 and 17 when both were reduced significantly. When previously fasted fish resumed feeding, plasma glucose and FFA, and hepatic glycogen levels rapidly returned to control values and weight gain resumed. No significant changes in plasma cortisol levels, related to feeding regime, were evident at any point during the study and there was no evidence that SL played an active role in the response to fasting. The results suggest that overwinter fasting may not represent a significant nutritional stressor to rainbow trout and that energy mobilisation during fasting may be achieved without the involvement of GH, cortisol or SL.     

Serum bone Gla protein (BGP) during treatment of hyperthyroidism and hypothyroidism. A longitudinal study

Serum bone gamma-carboxyglutamic acid-containing (Gla) protein (BGP) was measured before and with initially 2 weeks, later 4-8 weeks intervals for 20-58 weeks during treatment of patients with hyperthyroidism (n = 10) and hypothyroidism (n = 4). Biochemical euthyroidism was obtained in the hyperthyroid patients after a median of 3 weeks (range 1-8 weeks), and in the hypothyroid patients after a median of 17 weeks (range 10-27 weeks). Serum BGP levels closely followed the thyroid state, being high respectively low in the hyperthyroid and hypothyroid state and reaching a stable plateau just at the time biochemical euthyroidism was obtained. These data suggest that osteoblastic activity is enhanced in hyperthyroidism and reduced in hypothyroidism, and that normalization occurs in close conjunction with the normalization of the thyroid state, without any delay, indicating a direct effect on the function of the excisting osteoblasts by the thyroid hormones.     

Tesaglitazar, a dual PPAR{alpha}/{gamma} agonist, ameliorates glucose and lipid intolerance in obese Zucker rats

Insulin resistance, impaired glucose tolerance, high circulating levels of free fatty acids (FFA), and postprandial hyperlipidemia are associated with the metabolic syndrome, which has been linked to increased risk of cardiovascular disease. We studied the metabolic responses to an oral glucose/triglyceride (TG) (1.7/2.0 g/kg lean body mass) load in three groups of conscious 7-h fasted Zucker rats: lean healthy controls, obese insulin-resistant/dyslipidemic controls, and obese rats treated with the dual peroxisome proliferator-activated receptor alpha/gamma agonist, tesaglitazar, 3 mumol.kg(-1).day(-1) for 4 wk. Untreated obese Zucker rats displayed marked insulin resistance, as well as glucose and lipid intolerance in response to the glucose/TG load. The 2-h postload area under the curve values were greater for glucose (+19%), insulin (+849%), FFA (+53%), and TG (+413%) compared with untreated lean controls. Treatment with tesaglitazar lowered fasting plasma glucose, improved glucose tolerance, substantially reduced fasting and postload insulin levels, and markedly lowered fasting TG and improved lipid tolerance. Fasting FFA were not affected, but postprandial FFA suppression was restored to levels seen in lean controls. Mechanisms of tesaglitazar-induced lowering of plasma TG were studied separately using the Triton WR1339 method. In anesthetized, 5-h fasted, obese Zucker rats, tesaglitazar reduced hepatic TG secretion by 47%, increased plasma TG clearance by 490%, and reduced very low-density lipoprotein (VLDL) apolipoprotein CIII content by 86%, compared with obese controls. In conclusion, the glucose/lipid tolerance test in obese Zucker rats appears to be a useful model of the metabolic syndrome that can be used to evaluate therapeutic effects on impaired postprandial glucose and lipid metabolism. The present work demonstrates that tesaglitazar ameliorates these abnormalities and enhances insulin sensitivity in this animal model.     

Basal and stimulated hyperinsulinemia in obesity: relationship to adipose-cell size.

In 17 non-selected, non-hyperlipoproteinemic subjects without overt diabetes both adipose tissue biopsy and an oral glucose tolerance test (50 g) were performed. All persons were weight-stable at the time of investigation. A significant correlation between fasting insulin concentration and mean adipocyte volume was observed, whereas no correlation existed between ideal body weight index and fasting insulin level. Persons with larger adipocytes had elevated basal insulin levels as well as higher and longer lasting increments following the glucose challenge. They also exhibited significantly higher mean glucose levels during the OGTT. When these patients were matched for glucose tolerance with the subgroup having smaller mean adipocyte volumes, the difference in insulin levels was still demonstrable. This study underlines the importance of adipose-cell enlargement regulating basal and stimulated insulin output.     

Plasma free fatty acid transport during prolonged glucose consumption and its relationship to plasma triglyceride fatty acids in man

Plasma free fatty acid (FFA) transport in human subjects has been studied during the course of prolonged ingestion of different amounts of glucose. Compared with the fasting state, hypocaloric glucose intake resulted in marked suppression of net transport of FFA with no change in (fractional) turnover rate. There was no further suppression of net transport of FFA when the intake was increased to isocaloric or hypercaloric levels, but there was a significant increase in the (fractional) turnover rate, indicating an enhancement of clearance mechanisms. During the 20-24-hr period of fasting after isocaloric glucose consumption, the (fractional) turnover rate quickly fell to that found in the fasting individual, whereas net transport remained suppressed for much longer. This suggested that ingestion of glucose maintains an influence on lipolysis longer than on esterification. During this period of fasting after glucose administration, the contribution of plasma FFA to circulating triglyceride fatty acids increased with time and was positively and significantly correlated with the changes in the net transport of plasma FFA.     

Hyperinsulinemia in obesity: significance of adipose tissue cell size

Adipose tissue biopsy and an oral glucose tolerance test (OGTT) (50 g) were performed in 17 non-hyperlipoproteinemic subjects without overt diabetes mellitus. All the persons were weight stable at the time of investigation. A significant correlation between fasting insulin concentration and the mean adipocyte size was observed, whereas no correlation was noted between the ideal body weight index and fasting insulin level. Persons with larger adipocytes had elevated insulin levels as well as higher and longer lasting increments following the glucose challenge. They also exhibited significantly higher mean glucose levels during the OGTT. When these patients were matched for glucose tolerance with the subgroup having smaller mean adipocyte sizes, the difference in insulin levels was still demonstrable. The importance of adipose cell enlargement regulating basal and stimulated insulin output is underlined.     

Metabolic effects of chronic prazosin treatment

The effects of chronic (3 mg/day for 1 week) administration of the vasodilator drug prazosin on several metabolic and endocrine variables were evaluated in 12 hypertensive patients, 6 with normal and 6 with abnormal oral glucose tolerance test (OGTT). After 1 week prazosin treatment there were no significant modifications in fasting plasma glucose, serum free fatty acids (FFA), cholesterol, triglycerides, insulin (IRI), growth hormone (GH), prolactin (PRL) and gastrin levels; oral glucose tolerance and IRI response to glucose were unchanged in normal subjects, while in chemical diabetics there was a significant improvement in glucose tolerance and a slight increse in IRI secretion. Therefore, the untoward metabolic effects of acute prazosin administration, i.e. increased plasma glucose and serum FFA, are not sustained during chronic treatment, which may even improve glucose metabolism in diabetic patients.     

Effects of age on fasting-induced changes in insulin, glucose, urea nitrogen, and free fatty acids in sera of sheep

The hypothesis that prepubertal ewe lambs are metabolically different from postpubertal ewes was tested. Ovariectomized ewes (4 years of age; n = 4) and lambs (6 months of age; n = 4) were fasted for 72 hr. Serum concentrations of insulin, glucose, urea nitrogen, and free fatty acids (FFA) were measured in blood samples taken at 6-hr intervals between 30 hr before and 72 hr after feed removal. Serum concentrations of urea nitrogen and glucose were not different (P greater than 0.20) between age groups before fasting. Serum concentrations of insulin in ewes increased toward the end of the prefast period whereas those in lambs did not (age x time, P less than 0.01). Serum concentrations of FFA in ewes tended to be lower (P less than 0.07) than those in lambs prior to fasting. During fasting, concentrations of insulin decreased (P less than 0.02) over time in ewes and lambs and did so in a similar manner (age x time, P greater than 0.70). Urea nitrogen increased (P less than 0.0001) in both fasted ewes and fasted lambs in a comparable manner (age x time, P greater than 0.20). Concentrations of glucose during fasting were not significantly affected (P greater than 0.90) by age. There was a tendency (P = 0.08) for concentrations of glucose to change over time but the pattern did not appear to be related to fasting. During fasting, concentrations of FFA tended to be higher (P less than 0.07) in lambs than in ewes and increased (P less than 0.0001) in both groups in a similar fashion (age x time, P greater than 0.10). The findings herein suggest that turnover of FFA in lambs may be slightly greater than that in ewes during the fed and fasted states.     

Measurement of serum thyroxine binding prealbumin in various thyroidal states by radioimmunoassay

Serum thyroxine binding prealbumin (TBPA) levels in various thyroidal states were examined by radioimmunoassay (RIA). This technique is highly sensitive, accurate and reproducible. The normal mean (+/- 2SD) level of serum TBPA is 26.9 +/- 8.0 mg/dl (29.4 +/- 5.2 in men and 24.9 +/- 7.6 mg/dl in women). Serum TBPA levels in pregnant women were significantly lower than in normal females (P less than 0.05). Serum TBPA levels in patients with untreated hyperthyroidism were 12.9 +/- 4.0 mg/dl (mean +/- SD) and in patients with untreated hypothyroidism were 25.2 +/- 4.7 mg/dl (mean +/- SD). The mean TBPA concentrations in untreated hyperthyroidism were significantly lower than that for normal population (P less than 0.01), but untreated hypothyroidism was almost within normal range. The changes in TBPA levels in hyperthyroidism and hypothyroidism were similar to those in TBG levels. In untreated hyper- and hypothyroidism, restoration to euthyroidism by treatment was uniformly accompanied by a normalization of serum TBPA and TBG levels. A negative correlation between serum thyroid hormone binding protein (TBG and TBPA) and free thyroxine was observed in patients with hyperthyroidism. The coefficient of correlation between TBPA and free thyroxine was -0.80 (P less than 0.01) and between TBG and free thyroxine -0.58 (P less than 0.01). From these experiments it appears that not only TBG but also TBPA may play an important role in the regulation of the free thyroxine concentration in response to various thyroidal states.     

Experimental hyperthyroidism does not induce hepatic insulin resistance in the miniature pig.

The effect of hypo- and hyper-thyroidism on insulin-mediated alterations in tracer-determined glucose kinetics and the arterial concentration of gluconeogenic precursors were investigated in 24 h-starved conscious unrestrained miniature pigs. Hyperinsulinaemia (about 40 microunits/ml) decreased blood glucose and, transiently, glucose output at unaltered glucose utilization in all thyroid states: this effect was pronounced in hyperthyroid (-50%) and less in hypothyroid pigs (-25%) compared with euthyroid controls (-35%). We conclude that moderate experimental hyperthyroidism does not induce hepatic insulin resistance, whereas hypothyroidism slightly impairs insulin action with respect to the regulation of glucose output.     

Effect of glucose infusion on venous blood levels of immunoreactive proinsulin activity, insulin activity and fat parameters in healthy and protodiabetic subjects.

Concentrations of immunoreactive insulin activity (IRI) and proinsulin activity (IRP), blood glucose, free fatty acids (FFA), glycerol, cholesterol, triglycerides were analyzed in 140 subjects suspect of protodiabetes and 50 healthy persons before, during and after a glucose infusion test (GIT). The protodiabetic subjects were classified into normweight, overweight, obese, hyperlipemic groups with diet or with Regadrin therapy and each of them subdivided into such with normal and such with pathological carbohydrate tolerance. Norm- and overweight subjects with asymptomatic diabetes were characterized by a significant reduction of insulin secretion during both phases. Obese patients with or without hyperlipoproteinemia demonstrated an increased IRI reaction during the late phase of secretion. Carbohydrate intolerance was associated with an enhancement of basal triglyceride levels and a reduced depression of glycerol and FFA during the GIT. There were no differences in fasting or reactive IRP concentrations between healthy and protodiabetic subjects with normal carbohydrate tolerance. In asymptomatic diabetes the IRP levels were increased during the late secretion phase, but the percentage of IRP in total IRI was normal or--in existing high response--significantly reduced in comparison to norm response. The results do not support an enhanced IRP secretion as the cause of carbohydrate intolerance.     

Infusion of nicotinic acid stimulates leucine oxidation and inhibits protein synthesis in pigs before and during a meal

Leucine metabolism was measured isotopically in immature female pigs to assess the effect of acute infusions of nicotinic acid (NA) on leucine kinetics in both the fed and fasting states. After an overnight fast, immature pigs were infused with 3H-alpha-ketoisocaproate (KIC) and 14C-leucine. After a 2-hour equilibration period, an infusion of either saline or 0.4 mg/kg.min of NA was begun. NA caused a decrease in plasma glucose and an increase in plasma glucagon. During the fasting period, NA increased KIC oxidation 2-fold over controls. After feeding, plasma free fatty acids (FFA) in both groups were equivalent, but KIC oxidation was still approximately 80% higher in NA-infused animals. In addition, NA stimulated proteolysis and inhibited protein synthesis during the meal. Because plasma FFA concentrations were equal during the fed period, it is unlikely that changes in FFA concentrations are responsible for the changes in leucine metabolism observal during NA infusion.     

Nocturnal free fatty acids are uniquely elevated in the longitudinal development of diet-induced insulin resistance and hyperinsulinemia

Obesity is strongly associated with hyperinsulinemia and insulin resistance, both primary risk factors for type 2 diabetes. It has been thought that increased fasting free fatty acids (FFA) may be responsible for the development of insulin resistance during obesity, causing an increase in plasma glucose levels, which would then signal for compensatory hyperinsulinemia. But when obesity is induced by fat feeding in the dog model, there is development of insulin resistance and a marked increase in fasting insulin despite constant fasting FFA and glucose. We examined the 24-h plasma profiles of FFA, glucose, and other hormones to observe any potential longitudinal postprandial or nocturnal alterations that could lead to both insulin resistance and compensatory hyperinsulinemia induced by a high-fat diet in eight normal dogs. We found that after 6 wk of a high-fat, hypercaloric diet, there was development of significant insulin resistance and hyperinsulinemia as well as accumulation of both subcutaneous and visceral fat without a change in either fasting glucose or postprandial glucose. Moreover, although there was no change in fasting FFA, there was a highly significant increase in the nocturnal levels of FFA that occurred as a result of fat feeding. Thus enhanced nocturnal FFA, but not glucose, may be responsible for development of insulin resistance and fasting hyperinsulinemia in the fat-fed dog model.     

Sensitivity of the soleus muscle to insulin in resting and exercising rats with experimental hypo- and hyper-thyroidism.

1. The effects of hypothyroidism (caused by surgical thyroidectomy followed by treatment for 1 month with propylthiouracil) and of hyperthyroidism [induced by subcutaneous administration of L-tri-iodothyronine (T3)] on glucose tolerance and skeletal-muscle sensitivity to insulin were examined in rats. Glucose tolerance was estimated during 2 h after subcutaneous glucose injection (1 g/kg body wt.). The sensitivity of the soleus muscle to insulin was studied in vitro in sedentary and acutely exercised animals. 2. Glucose tolerance was impaired in both hypothyroid and hyperthyroid rats in comparison with euthyroid controls. 3. In the soleus muscle, responsiveness of the rate of lactate formation to insulin was abolished in hypothyroid rats, whereas the sensitivity of the rate of glycogen synthesis to insulin was unchanged. In hyperthyroid animals, opposite changes were found, i.e. responsiveness of the rate of glycogen synthesis was inhibited and the sensitivity of the rate of lactate production did not differ from that in control sedentary rats. 4. A single bout of exercise for 30 min potentiated the stimulatory effect of insulin on lactate formation in hyperthyroid rats and on glycogen synthesis in hypothyroid animals. 5. The data suggest that thyroid hormones exert an interactive effect with insulin in skeletal muscle. This is likely to be at the post-receptor level, inhibiting the effect of insulin on glycogen synthesis and stimulating oxidative glucose utilization.     

Serum Metabolomic Patterns in Patients with Autoimmune Thyroid Disease

Objective: Autoimmune thyroid disease, including Graves disease (GD) and Hashimoto thyroiditis (HT), is one of the most common endocrine diseases. GD and HT are the main etiologies for hyperthyroidism and hypothyroidism, respectively. This study aimed to provide a metabolomic analysis of GD patients with hyperthyroidism and HT patients with hypothyroidism.Methods: This study investigated serum metabolomics in 43 GD patients with hyperthyroidism, 45 HT patients with hypothyroidism, and 52 age- and sex-matched healthy controls. The metabolomic data were analyzed by performing multivariate statistical analysis.Results: The 186 metabolites including amino acids, bile acids, free fatty acids, and lipids were identified in all participants. Multivariate models indicated systematic differences in the hyperthyroidism, hypothyroidism, and control groups. Compared to healthy controls, the 22 metabolites in the hyperthyroidism group and the 17 metabolites in the hypothyroidism group were significantly changed. Pathway analysis showed that hyperthyroidism had a significant impact on arginine and proline metabolism and aminoacyl-transfer ribonucleic acid biosynthesis, while hypothyroidism had a significant impact on alanine, aspartate, and glutamate metabolism.Conclusion: The serum metabolomic pattern changes in patients with autoimmune thyroid dysfunction.Abbreviations: BMI = body mass index; CA = cholic acid; CDCA = chenodeoxycholic acid; DCA = deoxycholic acid; FBG = fasting plasma glucose; FINS = fasting plasma insulin; FT3 = free triiodothyronine; FT4 = free thyroxine; GD = Graves disease; GDCA = glycodeoxycholic acid; HDL-C = high-density lipoprotein cholesterol; HOMA-IR = homeostasis model assessment of insulin resistance; HT = Hashimoto thyroiditis; LDL-C = low-density lipoprotein cholesterol; PC = phosphatidylcholine; PCA = principal component analysis; PLS-DA = partial least squares discriminant analysis; SM = sphingomyelin; TBA = total bile acid; TC = total cholesterol; TG = triglyceride; TSH = thyrotropin; VIP = variable influences on projection