Association Of Dynamic Changes In Metabolic Factors With Risk Of Diabetes And Cardiovascular Complication

Diabetes has become a major public health threat to people’s health and socioeconomic development in China.Cardiovascular disease(CVD)is the most common complication of type 2 diabetes and a leading cause of death and disability in people with diabetes.Thus,it is essential to prevent diabetes and cardiovascular complication.During the development of diabetes and its complications,dynamic changes in various metabolic factors such as blood glucose,insulin,blood pressure,blood lipid,and weight are important phenotypic characteristics.However,large between-person variations exist in populations,and most of the previous studies only evaluated the risk of developing diabetes and its complications based on a single measurement.Currently,there is a lack of large cohort studies that use repeated measures over time to reveal the association of dynamic changes in metabolic factors with risk of diabetes and cardiovascular complication.Recently,few studies with repeated measures have investigated dynamic changes in metabolic factors before diabetes or CVD.However,some limitations still needed to be noticed.First,some studies have a small sample size or number of cases,limited measures,or long follow-up time intervals.Second,most studies were conducted in Western countries,while evidence from the Chinese population was rather scarce.Third,some studies only focused on the changes of certain metabolic factors,without a comprehensive consideration of the changes of multiple factors throughout the development of diabetes.Fourth,present studies mainly focused on dynamic changes in metabolic factors before the diagnosis of diabetes,while no study has explored the dynamic changes in metabolic factors during the development of cardiovascular complication exclusively in diabetic patients.In view of this,this study was based on the Kailuan Study to evaluate the dynamic changes in metabolic factors.The Kailuan Study is a large prospective dynamic cohort,which recruited active and retired employees from the Kailuan Group(Tangshan,Hebei Province).All participants underwent standardized questionnaire assessments,physical examinations,and laboratory tests in the Kailuan General Hospital and its 10 affiliated hospitals.The study was launched in 2006 to 2007(cycle 1)and more than 100 thousand participants completed the first health examination.Thereafter,health examinations were conducted biennially,and approximately 10 to 20 thousand participants were newly recruited in 2008 to 2009(cycle 2),2010 to 2011(cycle 3),2012 to 2013(cycle 4),2014 to 2015(cycle 5),2016 to 2017(cycle 6),and 2018 to 2019(cycle 7),respectively.Until now,the Kailuan Study has collected complete data from seven health examinations.Due to the high proportion of undiagnosed diabetes in the Chinese population(over 50%),diagnosis of diabetes through the hospital medical record system would seriously underestimate the incidence of diabetes,therefore information on diabetes diagnosis in this study was obtained through questionnaires and laboratory test information at each health examination.The occurrence of CVD was acquired by linking to the hospital medical record system.This study used repeated measurements(up to seven times)of participants to characterize the dynamic changes of 16 metabolic factors(including fasting glucose,surrogate index of insulin resistance or insulin sensitivity,blood lipid,blood pressure,heart rate,obesity indicator,liver function indicator,and inflammatory factor)during the development of diabetes and cardiovascular complication,and to evaluate the association of different trajectories of metabolic factors with risk of diabetes and cardiovascular complication.We aimed to provide new insights for timely prevention,early diagnosis and personalized intervention of diabetes and cardiovascular complication.Part 1 Association of dynamic changes in metabolic factors with risk of diabetesObjective: To characterize the dynamic changes in metabolic factors before diagnosis of diabetes,explore the critical period during the development of diabetes,and evaluate the association between different trajectories of metabolic factors and risk of diabetes.Methods: Considering that the information of diabetes diagnosis were collected through follow-up health examinations in this study,a total of 125367 participants without prevalent diabetes who participated in cycles 1 to 6 were selected,and followed until cycle 7.Diabetes was defined as fasting blood glucose of 7.0 mmol/L or higher,self-report of a physician diagnosis,or use of antidiabetic medication.The time when a participant first completed the health examination served as the baseline.For those who developed diabetes,the time of diabetes diagnosis was used as the end of follow-up,and for those who did not develop diabetes,the time of last participation in the health examination was used as the end of follow-up.Participants were divided into two groups according to their diabetes status during the follow-up: those who did and did not develop diabetes.Multivariableadjusted piecewise growth mixture models were used to fit the dynamic changes and compare differences of each metabolic factor in the two groups,and to explore the critical period when levels of metabolic factors changed dramatically.Furthermore,59982 participants who attended three consecutive health examinations were selected,and latent class growth models were used to classify changes of metabolic factors into different trajectory groups;multivariable-adjusted Cox proportional hazard regression models were then used to evaluate the association of different trajectory groups with the risk of diabetes.Results: The mean age of 125367 participants was 48.1 years,and during 14 years of follow-up,16935(13.5%)participants were newly diagnosed with diabetes.Most metabolic factors in participants who did not develop diabetes remained relatively stable over the follow-up period.During the 14 to 3 years before the diabetes diagnosis,changing patterns of most metabolic factors were similar to that of those who did not develop diabetes;levels of insulin sensitivity and high-density lipoprotein cholesterol remained consistently lower,and 14 other metabolic factors were consistently higher.During the 3 years before the diabetes diagnosis,multiple metabolic factors experienced abrupt changes,among which 7 metabolic factors including fasting glucose,triglyceride-glucose index,total cholesterol,triglyceride,heart rate,alanine aminotransferase,and hypersensitive C-reactive protein increased in nonlinear manners;4 metabolic factors including systolic blood pressure,pulse pressure,low-density lipoprotein cholesterol,and visceral adiposity index increased linearly;1 metabolic factor(insulin sensitivity)decreased in a nonlinear manner;and 4 metabolic factors including high-density lipoprotein cholesterol,diastolic blood pressure,body mass index and waist circumference decreased linearly.For example,fasting glucose increased slightly over time(mean 0.010 mmol/L per year)in participants who did not develop diabetes,with concentrations of 5.11 mmol/L at 14 years before the end of follow-up and 5.26 mmol/L at the end of follow-up.In participants who developed diabetes,fasting glucose increased from 5.28 mmol/L at 14 years before diagnosis to 5.89 mmol/L at 3 years before diagnosis,and the mean increasing rate was 0.056 mmol/L per year,faster than that of those who did not develop diabetes(P <0.001);during the 3 years before the diabetes diagnosis,fasting glucose increased abruptly in a nonlinear manner,rising to 8.08 mmol/L at the time of diagnosis,2.82 mmol/L higher than that of those who did not develop diabetes(P <0.001).In addition,analysis of different trajectories of metabolic factors showed that compared with the low-stable trajectory group,the high-stable group of insulin sensitivity and high-density lipoprotein cholesterol were inversely associated with risk of diabetes,the corresponding hazard ratio(HRs)and 95% confidence interval(CIs)were 0.26(0.22,0.30)and 0.88(0.83,0.93),respectively.Different trajectories of the other 14 metabolic factors were positively associated with risk of diabetes,with HRs(95% CIs)ranging from 1.22(1.13,1.33)to 18.99(16.52,21.82)for the low-elevated,high-stable,or high-elevated group.Conclusions: We identified 3 years before diagnosis as the critical period when most metabolic factors experienced drastic changes.During the 14 years before diabetes diagnosis,all the 16 metabolic factors were already at more unfavorable levels and the changing patterns were significantly different.Different trajectories of metabolic factors were associated with the risk of developing diabetes.Part 2 Association of dynamic changes in metabolic factors with risk of cardiovascular complication in patients with newly diagnosed diabetesObjective: To characterize the dynamic changes in metabolic factors before the incidence of CVD in patients with newly diagnosed diabetes,explore the critical period,and evaluate the association between different trajectories of metabolic factors and risk of CVD.Methods: A total of 21798 patients with newly diagnosed diabetes who participated in cycles 1 to 7 were selected.The time when diabetes was diagnosed served as the baseline.Participants were followed up to the time of CVD incidence,death,or the end of update of CVD information(December 31,2021).Participants were divided into two groups according to whether CVD occurred during the follow-up: those who did and did not develop CVD.Piecewise growth mixture models were used to fit the dynamic changes and compare differences of each metabolic factor in the two groups,and to identify the critical period when levels of metabolic factors changed dramatically.Furthermore,12856 participants who attended three consecutive health examinations were selected,and latent class growth models were used to classify changes of metabolic factors into different trajectory groups;Cox proportional hazard regression models were then used to evaluate the association of different trajectory groups with the risk of CVD.Results: The mean age of 21798 participants was 55.7 years.During 15 years of follow-up,2105(9.7%)incident CVD cases were recorded.Compared with those who did not develop CVD,13 metabolic factors including insulin resistance,total cholesterol,blood pressure,obesity indicator,alanine aminotransferase and hypersensitive C-reactive protein were higher or the changing patterns were significantly different in those who developed CVD,while the levels and changing patterns of 3 metabolic factors including fasting glucose,insulin sensitivity,and high-density lipoprotein cholesterol were not significantly different.During the 15 years before the onset of CVD,systolic blood pressure increased nonlinearly,and the other 15 metabolic factors changed linearly: 6 metabolic factors including fasting glucose,triglyceride-glucose index,pulse pressure,heart rate,alanine aminotransferase and hypersensitive C-reactive protein increased linearly in both groups;1 metabolic factor(visceral adiposity index)decreased linearly in those without CVD and increased linearly before the onset of CVD;and 8 metabolic factors including insulin sensitivity,total cholesterol,triglyceride,low-density lipoprotein cholesterol,high-density lipoprotein cholesterol,diastolic blood pressure,body mass index and waist circumference decreased linearly in both groups.Compared with the low-stable group,the high-stable group of insulin sensitivity was inversely associated with the risk of CVD,with an HR(95% CI)of 0.65(0.49,0.85);trajectories of 12 others metabolic factors were positively associated with risk of CVD,with HRs(95% CIs)ranging from 1.27(1.10,1.47)to 2.66(2.21,3.20)for the low-elevated,moderate-elevated,high-stable,or high-elevated group.Conclusions: During the 15 years before the onset of CVD in patients with newly diagnosed diabetes,systolic blood pressure increased in a nonlinear manner,and the other metabolic factors changed linearly;Before the incidence of CVD,most metabolic factors were in consistently higher levels or showed significantly different changing patterns compared with those who did not develop CVD.Trajectories of multiple metabolic factors were associated with the risk of CVD.