Association Of Thyrotropin And Lipid Profile In Patients With Newly Diagnosed Asymptomatic Coronary Heart Disease

BackgroundTSH is the major regulator of the morphologic and functional states of the thyroid gland, which is synthesized and secreted by the thyrotrophs in the anterior pituitary. TSH binds with its specific cell membrane receptor-TSHR expressed on the membrane of thyroid follicular cells and leads to the stimulation of secondary messenger signal pathways predominantly mediated by cAMP. TSH secretion is stimulated by thyrotropin releasing hormone(TSHR) and inhibited by the circulating thyroid hormone (TH) in a classical negative feed-back loop. Physiological roles of TSH mainly lie in the stimulation of differentiated thyroid functions, such as iodine uptake and organification, production and release of TH from the gland, and promotion of thyroid growth. TSH responds significantly and precisely to minor changes of circulating TH concentrations, and it is regarded as the most sensitive criteria in reflecting the changes of the function of hypothalamus-pituitary-thyroid axis.Among the thyroid dysfunction-related diseases, dyslipidemia and coronary heart disease(CHD) have been the points of interests. Researches indicated that the association exists not only between overt thyroid dysfunction and the onset and development of dyslipidemia and CHD, but also between subclinical hypothyroidism(SCH) and the above abnormalities, although the results were inconsistent. SCH is a minor thyroid dysfunction characterized by the normal levels of TH and elevated levels of TSH. More importantly, in recent years, TSHR has been found to be widely expressed on extra-thyroidal tissues and organs, including white and brown adipocytes, renal cells, bone marrow and even hepatocytes. Some researches indicated that there existed the TSH-TSHR interaction in the above cells under treatment of TSH, which suggested that the existence of target organs of TSH outside the thyroid gland. More importantly, some research in human primary hepatocyte, rat liver tissue and rat hepatic cell line indicated that the hepatic expression and enzyme activity of hydroxy methyl glutarylcoenzyme A reductase(HMGCR) was stimulated under the treatment of TSH, which indicated a potential role of TSH in regulation of cholesterol synthesis in liver.Thus, as clinical investigation part of the project which was supported by National Natural Science Foundation of China (30971409), Natural Science Foundation (ZR2009CZ009) and the international cooperation grant (2011) of Shandong Province of China, we preliminarily planed to investigate the possible association between serum TSH level and lipid profile in patients with newly diagnosed asymptomatic CHD. Then after excluding smoking which is one of the important confounding factors that influencing both TSH levels and lipid parameters, we further performed analysis regarding the association between TSH within normal range and lipid parameters to get a more detailed understanding in this field.Objective1. To investigate the distribution of thyroid function status in patients with newly diagnosed asymptomatic CHD and compare the clinical parameters between subjects with thyroid dysfunctions and euthyroid ones. To obtain the preliminary relationship between TSH and lipid profiles in the euthyroid patients with newly diagnosed asymptomatic CHD.2. To investigate the distribution of thyroid function status in non-smokers with newly diagnosed asymptomatic CHD and compare the clinical parameters between subjects with thyroid dysfunctions and euthyroid ones.3. To explore the relationship between serum TSH level within normal range and lipid profiles in non-smokers with newly diagnosed asymptomatic CHD.4. Make a preliminary probe into the association of TSH within normal range and blood pressure.Subjects and Methods1. Subjects:(1) This retrospective study recruited921patients older than45years of age, who had been newly diagnosed with asymptomatic CHD confirmed by coronary angiography at the health center of Qianfoshan Hospital or Shandong Provincial Hospital (Shandong, China) between January,2004and December,2010. After excluding the subjects according to the exclusion criteria, the remaining689cases were collected as the preliminary study population. Based on the thyroid function assay, these patients were divided into hypothyroidism group, SCH group, euthyroid group, subclinical hyperthyroidism group and hyperthyroidism group.(2)540euthyroid patients selected from the above study population were adopted to analyze the association of TSH within normal range and lipid profiles by path model analysis.(3) In second part of the study,521non-smokers with newly diagnosed asymptomatic CHD were obtained from the first part by excluding past and current smokers. These patients were divided according to thyroid function and the clinical characteristics were compared between euthyroid patients and those with SCH and hypothyroidism. In the406euthyroid patients, the relationship between lipid parameters and TSH levels was investigated either as a whole or as groups after dividing by either lipid profiles or different TSH categories within normal range.2.Detected index:Data were obtained from medical records. All the patients were stable at the moment of blood samples collection. The levels of FT3, FT4, TSH, anti-thyroglobulin antibody (Anti-TGAb) and anti-thyroid peroxidase antibody (Anti-TPOAb) were measured by using an electrochemiluminescence immunoassay. The lipid profiles including total cholesterol (TC), triglyceride (TG), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol(HDL-C) and other laboratory data including fasting blood glucose (FBG) and uric acid (UA) level were measured using an automatic biochemistry analyzer. Non-high density lipoprotein cholesterol (non-HDL-C) was calculated by subtracting HDL-C from TC. Systolic bold pressure (SBP) and diastolic blood pressure (DBP) was measured twice with a desk-model sphygmomanometer. The rest and exercise stress ECG were assessed.3. Stastical Analysis:Before statistical analysis, the normality of the distribution and the homogeneity of the variances were evaluated using Kolmogorov-Smirnov test and Levine's test, respectively. The log transformation of lipid parameters and blood pressure was performed to correct the skewed distribution. Statistical differences of the clinical characteristics were performed with ANOVA followed by a multiple-comparison test for subgroups by least significance difference (LSD). The difference of TSH within normal range between normal and abnormal lipid groups was tested by Mann-Whitney test. The relationships between gender, age, history of DM, history of hypertension, drinking, TSH, FT3, FT4, FBG, UA and the log-transformed lipid profile were assessed by simple correlation analysis. Stepwise multiple linear regression analysis, logistic regression analysis and general linear model analysis was performed to estimate the influence of the above parameters on the lipid profile. Path model analysis was used to evaluate the total, direct and indirect effects of TSH on lipid profiles. The level of statistical significance was set at p<0.05, except for the regression analysis, in which p<0.10was set as the cut-off point.Results 1. Distribution of the thyroid status and comparison of the general data in the689patients with newly diagnosed asymptomatic CHD(1)Most of the patients were euthyroid (78.4%), while the prevalence of thyroid dysfunction including subclinical ones was21.6%. SCH accounted for8.6%, while subclinical hyperthyroidism accounted for only1.5%. There were more patients with hypothyroidism than those with hyperthyroidism including the subclinical ones(14%vs.7.7%).(2)54cases in hypothyroidism(10males),66.92±10.50y old, the median of TC, TG, HDL-C and LDL-C were5.83,1.57,1.31and3.35mmol/L, respectively;59cases in SCH (20males),64.22±10.57y old, the median of TC, TG, HDL-C and LDL-C were5.34,1.41,1.34and2.95mmol/L, respectively;43cases in hyperthyroidism (18males),67.98±11.22y old, the median of TC, TG, HDL-C and LDL-C were4.87,1.37,1.12and3. OOmmol/L, respectively;10cases in subclinical hyperthyroidism (3males),7.60±8.18y old, the median of TC, TG, HDL-C and LDL-C were4.49,1.10,1.19and2.78mmol/L, respectively;540cases in euthyroidism (257males),64.85±10.81y old, the median of TC, TG, HDL-C and LDL-C were5.04,1.37,1.26and2.99mmol/L, respectively. Only the level of TC was significantly higher in hypothyroid patients than the euthyroid ones (p<0.01).2. Simple correlation of TSH with lipid profiles in689patients with newly diagnosed asymptomatic CHD and in the540euthyroid ones(1) In all the CHD patients, serum TSH level was positively correlated with log-transformed TC and HDL-C (r=0.206and0.092, p<0.01and p<0.05, respectively). TSH was not correlated with either log-transformed TG or LDL-C in this study population.(2) In the euthyroid CHD patients, serum TSH level was positively correlated with log-transformed TC and TG (r=0.202and0.096, p<0.01and p<0.05, respectively). TSH was not correlated with either log-transformed HDL-C or LDL-C in euthyroid patients. 3. Stepwise multiple linear regression analysis of the influence of TSH on lipid profiles in540euthyroid patients with newly diagnosed asymptomatic CHD(1) Serum TSH level was positively and linearly associated with Log TC (unstandarized coefficient=0.019, standardized B coefficient=0.191, p<0.001).(2) Serum TSH level was positively and linearly associated with Log TG (unstandarized coefficient=0.029, standardized B coefficients=0.113, p<0.05).(3) TSH was not associated with either Log HDL-C or Log LDL-C.4. Path model analysis of the effects of TSH on lipid profiles in540euthyroid patients with newly diagnosed asymptomatic CHD(1) The goodness of fit index (GFI) in the path model analysis>0.97in all the equations. The analysis based on the equations showed that the effects of TSH on Log TC and Log TG were significant (path coefficient=0.2028and0.1138, t=4.5037and2.5028, respectively).(2) The results of effects analysis showed that the total effects of TSH on TC, TG, LDL-C and HDL-C were as the followings:0.1936,0.1095,0.0808and-0.0391. The total effect of TSH on Log TC, Log TG and Log LDL-C was positive, while the total effect of TSH on Log HDL-C was negative. Further analysis indicated that the direct effect of TSH on Log TC, TG, LDL-C and HDL-C were0.2028,0.1138,0.0896and-0.0446, respectively. The indirect effect of TSH on Log TC, TG, LDL-C was negative, while the indirect effect of TSH on Log HDL-C was positive.5. Distribution of the thyroid status and comparison of the general clinical characteristics in521non-smokers with newly diagnosed asymptomatic CHD(1) Most of the non-smokers were euthyroid (77.9%), while the prevalence of thyroid dysfunction including subclinical ones was22.1%. SCH accounted for8.3%, while subclinicai hyperthyroidism accounted for only1.2%. There were more patients with hypothyroidism than those with hyperthyroidism including the subclinicai ones (14.4%vs.7.7%).(2) Only the level of TC was significantly higher in hypothyroid patients than that in the euthyroid ones (5.46vs.5.05mmol/L, p<0.05). TC in the SCH was also higher than that in euthyroid patients (5.36vs.5.05mmol/L), but did not reach statistical significance. The other lipid parameters did not differ significantly between SCH and euthyroid ones. No significant difference was found in both SBP and DBP between SCH and euthyroid controls.6. Stepwise multiple regression analysis of TSH on lipid profiles in406euthyroid non-smokers with newly diagnosed asymptomatic CHDThe results showed that within the normal range, TSH level was positively and linearly associated with Log TC, Log TG and Log non-HDL-C (standardized regression coefficient=0.017,0.027and0.026,p<0.05in all) in this study population.7. Relationship of blood pressure and serum TSH levels in406euthyroid non-smokers with newly diagnosed asymptomatic CHDRegression analysis showed that TSH had no significant effect on both Log SBP and Log DBP, and only the history of hypertension entered the regression equation. The standardized coefficient of the history of hypertension in the regression analysis of Log SBP was0.480(p<0.001), and that of Log DBP was0.322(p<0.001)8. Comparison of general characteristics, lipid profiles and blood pressure in406euthyroid non-smokers with newly diagnosed asymptomatic CHD according to different TSH categories(1) General data of patients in different TSH categories within the normal range of TSH were as the followings:79cases in G1(TSH0.3-0.99mIU/L,40males),67.04±10.63yold;135cases in G2(TSH1.0-1.89mIU/L, 69males),64.59±10.15y old;78cases in G3(TSH1.9-2.49mIU/L,41males),63.56±11.32y old;114cases inG4(TSH2.5-4.8mIU/L,37males),63.20±10.72y old. In analysis of clinical characteristics, only FBG was significantly lower in G2and G3when compared with those in G1(5.53±1.38vs.6.26±2.32,p<0.05;5.49±1.22vs.6.26±2.32,/p<0.01, respectively).(2) Setting G1as the control group, we compared the log-transformed lipid parameters in different TSH categories. Log TC was significantly higher in G2, G3and G4when comparing with G1(0.70±0.09,0.70±0.08and0.71±0.09vs.0.66±0.10mmol/L, p<0.01,<0.05and<0.01, respectively). Log non-HDL-C was also significantly higher in G2, G3and G4when comparing with G1(0.57±0.10,0.56±0.13and0.58±0.13vs.0.51±0.17mmol/L,p<0.01,<0.05and<0.01, respectively). Log TG was significantly higher in G2, G3and G4when comparing with G1(0.17±0.22,0.18±0.24and0.19±0.24vs.0.09±0.26mmol/L, p<0.05,<0.05and<0.01, respectively). The difference of Log LDL-C was only significantly between G2and G1(0.48±0.11vs.0.44±0.10mmol/L, p<0.05). There was no significant difference in the level of Log HDL-C among the four different categories of TSH within the normal range.(3) Neither SBP nor DBP differed significantly among the four categories according to TSH level within normal range in the euthyroid non-smokers with newly diagnosed CHD.9. Comparison of the prevalence of dyslipidemia in406euthyroid non-smokers with newly diagnosed asymptomatic CHD according to different TSH categoriesThe overall prevalence of hypercholesterolemia, hypertriglyceridemia, high LDL-C and low HDL-C were14.3%,33.5%,9.1%and5.7%, respectively. After dividing the patients with TSH in the normal range into four different categories, the prevalence of hypercholesterolemia was significantly higher in G2, G3and G4as compared with G1(15.6%,12.8%,21.1%vs.3.8%,p<0.01,<0.05and<0.01, respectively). The prevalence of hypertriglyceridemia was significantly higher in G2, G3and G4as compared with G1(34.8%,42.3%,36.8%vs.17.9%,p<0.05,<0.01and<0.01, respectively). The prevalence of abnormal high LDL-C was only significantly higher in G2and G3as compared with G1(11.3%and14.1%vs.2.5%, p<0.05and<0.01,respectively). The prevalence of abnormal low HDL-C was not significantly different among the four TSH categories.10. Comparison of the TSH levels in406euthyroid non-smokers with newly diagnosed asymptomatic CHD according to different lipid groupsBy Mann-Whitney test, the results showed that the median of TSH was significantly higher in hypercholesterolemia group as compared with that in the normal TC group (2.12Vs.1.76mmol/L,p<0.01). Meanwhile, Significant difference existed in the median of TSH levels between the hypertriglyceridemic group and the normal TG group (2.03vs.1.73mmol/L, p<0.05). TSH levels were of no significant difference between abnormal and normal groups according to the levels of HDL-C and LDL-C.11. The odds ratio of dyslipidemia with respect to the serum TSH levels in406euthyroid non-smokes with newly diagnosed asymptomatic CHDLogistic regression analysis indicated that the odds ratio of hypercholesterolemia with respect to serum TSH level was1.640(p<0.05), and the odds ratio of hypertriglyceridemia with respect to serum TSH level was1.349(p<0.05). There was no significant influence of serum TSH level to the risk of onset of abnormal LDL-C and HDL-C.Conclusions1.In the clinical research in patients with newly diagnosed asymptomatic CHD, we revealed and verified that the thyroid stimulating hormone-TSH might participate in regulation of lipid metabolism and exert adverse effect on lipid profile in the upper limits of the normal range. Thus, this key finding might be instructive in the prevention and treatment of the CHD in patients with hypothyroidism (including SCH).2. In this study, path model analysis and regression analysis had been creatively used in combination in the analysis of the association of variables and lipid profile. We confirmed not only the association between serum TSH level and lipid parameters, but also the main influence of TSH on TC by decomposing the effects via path model analysis.3.In this study, we excluded the influence of an important confounding factor-smoking, which might exert effects on both lipid profile and TSH level. Thus, in the non-smokers with newly diagnosed asymptomatic CHD, we creatively found that even with the normal range, serum TSH level was positively and linearly associated with TC, TG and non-HDL-C. Furthermore, TSH in the upper limits of the normal range might act as a risk factor for hypercholesterolemia and hypertriglyceridemia.