Prediction Models For Metabolic Syndrome Containing Glucose Component Based On Glycosylation Biomarker And Environmental Factors

Background Metbolic syndrome is a common product of a group of risk factors including high blood pressure,high blood glucose,central obesity,low level of high-density lipoprotein and high triglyceride level.These risk factors are also called the five components of MS.Metabolic syndrome itself is also a risk factor for cardiovascular diseases and diabetes.Researches on MS now are mainly refered to the relationship between number of abnormal components and MS related diseases.For example,abnormal lycometabolism has been proved to be a cause of hyperuricemia.Forootan M reported that number of abnormal MS components might lead to rectal cancer or pancreatic cancer.Besides,a five-year long observation on 1384 wrokers in Taiwan showed that for those who had one abnormal MS component,the risk of getting MS increased 2 times compared with the normal group.The risk was even 7 times higher when there were 2 abnormal components.However,no observation on the process of high blood glucose to MS has been reported.Among the five components,abnormal blood glucose is a pre-clinic sign for diabetes.Altough it is cheap and convenient to make diagnosis of diabetes using fasting plasma glucose(FPG),yet once an individual is diagnosed to be a patient with diabetes,the cost will be quite high and the life quality would be quite low.Data from the World Health Organization showed that from 2005 to 2015 in China,economic loss caused by diabetes and its related cardiovascular diseases was up to 55.7 billion dollars.The prevalence of diabetes in China is now 9.70%,which is much higher than that of 2.60% in year 2002.High prevalence leads to great economic burden to all society.Under such circumstances,it is so important to conduct early stage prevention of high blood glucose,which has an important public health meaning.To realize early stage prevention of high blood glucose,proteins or molecules,also called biomarkers,that change before blood glucose increases must be located.In recent years,glycoproteins come into view of researchers.Glycoproteins evolve form glycosylation,a kind of posttranslational modification of proteins.Many glycoproteins have been found to be specific indicators for cancers as they play important roles in signal transformation,immune mediation,protein degradation and so on.Glycoproteins tend to change before their related diseases developed.As far as our knowledge is concerned,no research has reported the association between level of blood plasma glycoproteins and blood glucose,which is what we want to reveal in this reseach.Objective: To evaluate the five-year long risk of getting MS after two components of MS are impaired including blood glucose.To explore the association between plasma glycan proteins and blood glucose level.Methods: This study focus on those who did physical check up from January 2009 to Feburary 2016 in Beijing Xiaotangshan Hospital.Demographic characteristics like sex and age were collected,as well as information on the blood routine examination,such as blood pressure,blood glucose,blood triglyceride,waist circumference and high-density lipoprotein.Statistical description was conducted using density distribution diagram and frequency table.Chi-square test,Will-Coxon rank sum test and t test were applied separately to make between-group comparison according to data type,the inspection level was set to be 0.05.According to the result of sample size estimation,year 2009 and year 2010 were set as baseline while year 2014 and 2016 were selected as follow-up.Apriori Algorithm was used to find out the 5-year period risk of MS in people who had two abnormal components(with blood glucose included).Specifically,indicators of support,confidence and lift were used to evaluate the risk.Joint model was applied to observe the correlation of other three component under combination of two components.As to laboratory detection,subjects were carefully selected according to inclusion criteria and exclusion criteria.Then High Performance Liquid Chromatography and mass spectrometer were applied to finish qualitative and quantitative analysis of glycan proteins.Bayesian method was applied to diminish the batch effect of glycan proteins that generated during the detection process.All subjects were splited into two groups according to glucose level(below 5.6mmol/L is defined as normal).Wilcoxon rank sum test was applied to make comparison of plasma glycan level between two groups and logistic regression was then applied to find out the impact factors of blood glucose.Receiver Operating Characteristic curve was applied to test the prediction efficiency of each potential biomarker for blood glucose.10-fold cross validation method was then applied to verify the prediction efficacy of potential biomarkers.Results: Risk evaluation of MS in people with two abnormal components including high blood glucose.In this study,15651 subjects who did physical check up in the same hospital from January 2009 to Feburary 2016 were included.After 4877 people who had MS at baseline were excluded,10764 subjects were selected to do analysis.Among them there were 7258 male and 3506 female.During the follow up time,1885 people had MS in total.Strong rules exsited for all combination of high blood glucose and one other abnormal component to MS.Support,confidence and lift for people with high blood glucose and HBP were 0.042,0.335 and 1.314 respectively.For people with high blood glucose and high TG,the values were 0.038,0.377 and 1.483 respectively.For people with high blood glucose and low HDLC,the values were 0.029,0.386 and 1.517 respectively.For people with high blood glucose and abnormal WC,the values were 0.059,0.324 and 1.271 respectively.In detail,for people with high blood glucose and HBP,the risk of getting MS five years later is 33.50% and in this group of people,the risk increses 1.314 times when compared with the whole population.For people with high blood glucose and high TG,the risk of getting MS after five years is 37.70% and in this group of people,the risk increases 1.483 times when compared with the whole population.For people with high blood glucose and low HDLC,the risk of getting MS five years later is 38.60%,and in this group of people,the risk increases 1.517 times when compared with the whole population.For people with high blood glucose and abnormal WC,the risk of getting MS in five years is 32.40%,and in this cluster of people,the risk increases 1.271 times when compared with the whole population.In people with abnormal blood glucose and abnormal blood pressure,the correlation between WC and TG is getting larger and larger during the five years,which means they change in the same direction and control measures on both sides are needed at the same time.Yet the relationship between WC and HDLC didn’t fluctuate much.Correlation of TG and HDLC is increasing from-0.207 to-0.319,indicating that when TG increases,the HDLC level decreases,so they both need to be pay attention to even if only one of them changes.In people with abnormal blood glucose and abnormal triglyceride.The relationship of WC and BP decreases,as well as the relationship between BP and HDLC,yet correlation of WC and HDLC is getting larger from-0.303 to-0.351,they both need intervention.In people with abnormal blood glucose and abnormal high-density lipoprotein,correlation of BP and TG decreases,as well as relationship between WC and TG.But correlation of WC and BP reversed from negative relationship to positive relationship,so they change in the same direction and control measures are needed for both of them at the same time.In people with abnormal blood glucose and abnormal waist circumference,the increasing of TG is accompanied by the decline of HDLC,so both of them need to be taken control of at the same time.The relationship between BP and TG does not change much;Correlation of BP and HDLC reversed from positive to negative,they also change in the same direction.Smoking,drinking,exercising,work pressure and sleep were found to be impact factors of metabolic syndrome containing component of blood glucose.The odds ratio were 1.725(1.337-2.226),3.370(2.255-5.036),1.760(1.343-2.306),1.968(1.494-2.591)and 1.481(1.089-2.015)respectively.Association between blood glycans and glucose of metabolic syndrome.In this part,212 subjects were included according to the result of sample estimation.Male accouted for 46.70% percent while female accunted for 53.30% percent.Those aged below 30 years old took up 57.50% percent of all.Chi-square test revealed that no significant difference on sex exsited between two groups(p=0.074).Also no difference on age distribution exsited.Univariate analysis showed that significant difference was found on plasma glycans of GP5,GP8,GP11,GP12,DG3,Trisijalo,BADS,BAMS,G3 and A2(p < 0.05).Specifically,blood level of GP5 in high blood glucose group(Median=2.160)is higher than that of normal group(Median=1.970).Blood level of GP11,GP12,Trisijalo,BADS,G3 and A2 in abnormal blood glucose group is also higher than that of normal group.Blood level of GP8,DG3 and BAMS in abnormal blood glucose group was significantly lower than that of normal group. Logistic regression showed that GP5,GP11 and A2 are statistically significant with p value under 0.01.And after adjusting sex and age,GP5 was found to be the protective factor of MS blood glucose component with odds ratio of 0.257(95% CI: 0.098 – 0.678).GP11 and A2 was found to be risk factors of blood glucose with odds ratio of 2.495(95% CI: 1.014 – 6.135)and 2.655(95% CI: 1.042 – 6.765)respectively.So glycan proteins GP5,GP11 and A2 may be the potneial biomarkers for blood glucose prevention.With one unit increase of GP5,the risk of abnormal blood glucose decreases 0.257 times.With one unit rise of GP11 and A2,the risk of abnormal blood glucose increases 2.495 times and 2.655 times.Specifically,if we use glycan protenins obtained above to predict the risk of abnormal blood glucose,areas under roc curve are all relatively large.For GP5,GP11 and A2,the areas are 0.694,0.684 and 0.710 respectively.A2 has the best predicting ability.Yet when using all the three indicators to construct a prediction model,the area reaches the peak of 0.760.Conclusion: The risk of getting MS in people with high blood glucose and HBP after five years is 33.50% and the risk is 1.314 times of the normal population.In this group,the correlation between WC and TG is getting larger,as well as the correlation of TG and HDLC.In people with high blood glucose and high TG level,the risk is 37.70% and is 1.483 times higher than that of all population.In this group,the correlation of WC and HDLC is getting larger.In people with high blood glucose and low HDLC level,the risk is 38.60% and is 1.517 times higher than that of all population.In this group,the correlation of WC and BP is getting larger.For people with high blood glucose and abnormal WC,the risk is 32.40% and is 1.271 times higher than that of all population.In this group,the correlation of BP and HDLC is getting larger.Environment factors had an influence on MS when compared to health population.Enviromental factors had no influence on the type of MS.For the plasma glycoproteins,GP5,GP11 and A2 are the impact factors of blood glucose,they may be the potential biomarkers for prediction of high blood glucose.Good prediction efficiency of GP5,GP11 and A2 can be obtained as the areas under ROC curve are 0.694,0.684 and 0.710 respectively.When using the model constructed by all the three glycans to make prediction of abnormal blood glucose,a better area under ROC curve can be obtained as 0.760.For the Ig G glycoproteins,GP2 is the influential factor of blood glucose,yet the prediction efficacy is smaller than that of plasma glycoproteins(0.670).Results of 10-fold cross validation showed that good prediction efficiency was obtained.