Myocardial Alterations In Patients With Type 2 Diabetes Mellitus:A Cardiovascular Magnetic Resonance Study

Background and ObjectivesType 2 diabetes mellitus?T2DM?is the most comment chronic disease.T2DM patients are at increased risk for cardiovascular disease,which is the leading cause of T2DM related mortality and morbidity.T2DM patients are associated with myocardial extracellular matrix,structural and functional alterations and finally at increased risk for left ventricular?LV?heart failure with preserved ejection fraction?HFPEF?,even in the absence of hypertension and coronary artery disease.Once HFPEF occurs,T2DM patients would be with poor prognosis.So early,noninvasive and in vivo detection of myocardial alterations will contribute to early diagnosis,risk stratification and prognosis of T2DM-related myocardial injury.However,currently there is a lack of comprehensive research about T2DM related myocardial extracellular matrix,structural and functional alterations and those relationships with duration of diabetes.With high resolution,good repeatability and multi-parameters,cardiovascular magnetic resonance?CMR?can be used to assess myocardial extracellular matrix,structural and functional alterations.The extracellular volume?ECV?,derived from CMR native and post-contrast T1 mapping and hematocrit?HCT?,is a biomarker of in vivo assessing the extracellular space in the myocardium.ECV is calculated from the following equation,ECV=?1-HCT?*?1/T1myocardium post-contrast-1/T1myocardium native?/(1/T1blood post-contrast-1/T1blood _native).So the accuracy of ECV calculation can be significantly affected by the accuracy of HCT measurement.Previously,HCT measurement was obtained from venipuncture blood test?measured HCT,HCTm?.Recently,T1_blood-based synthetic HCT?HCTsyn?was introduced to replace HCTm,which is convenient and concise.But the accuracy of HCTsyn is still a subject of dispute,especially at 3.0 T CMR.So,firstly,the accuracy of HCTsyn needs to be explored,and then it can be used to detect the T2DM related myocardial extracellular matrix alterations.Additionally,CMR cine has been regarded as the golden standard for the assessment of cardiac function and structure.So,it can be used to detect the T2DM related myocardial structural and functional alterations.The aims of this study were:1)to test the accuracy and clinic values of T1_blood-based HCTsyn and synthetic ECV?ECVsyn?in 3.0 T CMR;2)to perform CMR T1 mapping and its derived ECV to quantify extracellular matrix expansion and its relationship with duration of diabetes;3)to perform CMR cine to quantify LV structural alterations and those relationships with duration of diabetes;4)to perform CMR cine to quantify LV and left atrium?LA?functional alterations and those relationships with duration of diabetes,to explore relationships among extracellular matrix expansion,LV structural,functional and LA functional alterations and to determine which index is most sensitive for detection of T2DM-related cardiac changes.Material and methodsThe present study contained the following four parts:1.The accuracy T1_blood-based HCTsyn in 3.0 T CMR.A total of 226 subjects with CMR T1 mapping and HCT measurement taken on the same day as the CMR scanning were retrospectively enrolled,comprising healthy controls?HC,n=45?,T2DM patients?n=60?,hypertrophic cardiomyopathy?HCM?patients?n=93?and 28 other patients.Those subjects were randomly split into derivation?n=121,comprising 18 HC,34 T2DM,49 HCM and 20other patients?and validation?n=105,comprising 27 HC,26 T2DM,44 HCM and 8 other patients?groups.Correlation of T1_blood with HCTm was established in the derivation group and used in both the derivation and the validation groups.HCTsyn was used to calculate ECVsyn and HCTm was used to calculate measured ECV?ECVm?.The relationships between the ECVsyn and ECVm values were separately explored in the two groups.In addition,the differences in ECVsyn and ECVm among the healthy controls,T2DM and HCM were separately explored in the two groups.2.T2DM related LV extracellular matrix expansion.Sixty T2DM patients and forty-five healthy controls were recruited between June 2015 and March 2018.Finally,fifty-three T2DM and thirty-eight controls were enrolled after exclusion of participants with poor images quality or whose papillary muscles were too small to measure.Based on the duration of diabetes,all T2DM patients were categorized as short term T2DM?duration of diabetes<5years,n=18?or long term T2DM?duration of diabetes>5 years,n=35?.All participants were performed T1 mapping to obtain the native and post-contrast T1 values of myocardium and blood and ECV was calculated accordingly.All participants were performed cine to obtain LV myocardial mass index?MMi?and to obtain extracellular matrix volume index?ECMVi?and cellular volume index?CVi?.The differences in ECV,MMi,ECMVi and CVi among controls,short term T2DM and long term T2DM were compared.3.T2DM related LV structural alterations.All participants were performed cine to obtain end diastolic volume index?EDVi?and mass to volume ratio?MVR?.Fractal dimension?FD?was obtained using fractal analysis on cine images to quantify endocardial remodeling.The differences in EDVi,MVR and FD among controls,short term T2DM and long term T2DM were compared.The relationships between LV extracellular matrix expansion and structural alterations were explored.4.T2DM related LV and LA functional alterations.Two-dimensional feature tracking was performed on four-chamber long-axis cine images and mid-ventricular short-axis cine images to obtain the LV global longitudinal,circumferential and radial strain,peak systolic and diastolic strain rate.The biplane area-length method was used to obtain LA total ejection fraction?EF?,passive EF and active EF.The differences in LA and LV function among controls,short term T2DM and long term T2DM were compared.The relationship among LV extracellular matrix expansion,LV structural and LA and LV functional alterations were explored.Uni-and multi-variable logistic regressions were used to determine the most sensitive index of T2DM-related myocardial alterations.Results1.The accuracy T1_blood-based HCTsyn in 3.0 T CMR.?1?T1_blood-based HCTsyn using data from our center1)Relationship between HCT and 1/T1_blood using data from our center.The regression equation was:HCTsyn=561.6*(1/T1_blood)+0.098?R²=0.19,p<0.001?in the derivation group.2)The relationship between HCTm and HCTsyn in the derivation and validation group.The HCTm was modestly correlated with the HCTsyn in the derivation group?Slope=1.00,R²=0.19,p<0.001?and in the validation group?Slope=0.93,R²=0.18,p<0.001?.Bland-Altman analysis showed that the difference between HCTm and HCTsyn was 0.0%?95%CI:-8.2 to 8.2%?in the derivation group and-0.2%?95%CI:-9.2 to 8.8%?in the validation group.HCTm was strongly correlated with the difference between the HCTsyn and HCTm in the derivation group?Slope=-0.8082,R²=0.81,p<0.001?and in the validation group?Slope=-0.8102,R²=0.80,p<0.001?.3)The relationship between ECVm and ECVsyn in the derivation and validation group.The ECVm was strongly correlated with the ECVsyn in the derivation group?Slope=0.98,R²=0.87,p<0.001?and in the validation group?Slope=0.94,R²=0.70,p<0.001?.Bland-Altman analysis showed that the difference between ECVm and ECVsyn was 0.0%?95%CI:-4.4 to 4.4%?in the derivation group and 0.1%?95%CI:-4.4 to 4.7%?in the validation group.ECVm was weakly correlated with the difference between the ECVsyn and ECVm in the derivation group?Slope=-0.11,R²=0.10,p<0.001?and modestly in the validation group?Slope=-0.25,R²=0.21,p<0.001?.4)Comparison of differences in ECVm and ECVsyn in groups with ECVm<30.5%and>30.5%of the derivation and validation group.In group with ECVm<30.5%of the derivation group,paired t test showed that ECVm was smaller than ECVsyn?p=0.019?,and in group with ECVm>30.5%of the derivation group,ECVm was larger than ECVsyn?p=0.044?.In group with ECVm<30.5%of the validation group,paired t test showed that ECVm was smaller than ECVsyn?p=0.006?,and in group with ECVm>30.5%of the validation group,ECVm was larger than ECVsyn?p=0.034?.5)Comparison of differences in ECVm and ECVsyn among HC,T2DM and HCM in the derivation group.The ECVm was larger in T2DM patients than that in HC?29.1±3.1%vs.26.4±2.4%,p=0.002?,whereas,the ECVsyn did not differ between T2DM and HC?28.3±2.9%vs.26.9±2.2%,p=0.064?.Compared with HC,HCM patients were associated with higher ECVsyn and ECVm of the mid-ventricle.6)Comparison of differences in ECVm and ECVsyn among HC,T2DM and HCM in the validation group.The ECVm was larger in T2DM patients than that in HC?28.6±2.9%vs.25.8±3.2%,p=0.002?,whereas,the ECVsyn did not differ between T2DM and HC?28.0±2.3%vs.26.7±2.6%,p=0.068?.Compared with HC,HCM patients were associated with higher ECVsyn and ECVm of the mid-ventricle.The ECVm of free wall was larger in HCM patients than that in HC,whereas,the ECVsyn of free wall did not differ between HCM and HC.7)Use of ECVsyn to categorize patients with abnormal ECV.Using our center's normal cut-off of 31.8%,the use of ECVsyn would lead to a 6–12%incorrect categorization of patients in the derivation groups and 12-25%in the validation groups.?2?T1_blood-based HCTsyn using equation from previously published paper1)Relationship between HCT and 1/T1_blood from previously published paper.The regression equation was:HCTsyn=869.7*(1/T1_blood)-0.071.2)The relationship between HCTm and HCTsyn in the derivation and validation group.The HCTm was modestly correlated with the HCTsyn in the derivation group?R²=0.19,p<0.001?and in the validation group?R²=0.18,p<0.001?.Bland-Altman analysis showed that the difference between HCTm and HCTsyn was-0.2%?95%CI:-8.7 to 8.3%?in the derivation group and-0.2%?95%CI:-9.6 to 9.3%?in the validation group.HCTm was strongly correlated with the difference between the HCTsyn and HCTm in the derivation group?R²=0.57,p<0.001?and in the validation group?R²=0.56,p<0.001?.3)The relationship between ECVm and ECVsyn in the derivation and validation group.The ECVm was strongly correlated with the ECVsyn in the derivation group?R²=0.86,p<0.001?and in the validation group?R²=0.67,p<0.001?.Bland-Altman analysis showed that the difference between ECVm and ECVsyn was 0.1%?95%CI:-4.5 to 4.7%?in the derivation group and 0.1%?95%CI:-4.7 to 4.9%?in the validation group.ECVm was weakly correlated with the difference between the ECVsyn and ECVm in the derivation group?R²=0.05,p=0.011?and modestly in the validation group?R²=0.18,p<0.001?.4)Comparison of differences in ECVm and ECVsyn in groups with ECVm<30.5%and>30.5%of the derivation and validation group.In group with ECVm<30.5%of the derivation group,paired t test showed that ECVm was smaller than ECVsyn?p=0.025?,and in group with ECVm>30.5%of the derivation group,ECVm was similar with ECVsyn?p=0.169?.In group with ECVm<30.5%of the validation group,paired t test showed that ECVm was smaller than ECVsyn?p=0.011?,and in group with ECVm>30.5%of the validation group,ECVm was larger than ECVsyn?p=0.033?.5)Comparison of differences in ECVm and ECVsyn among HC,T2DM and HCM in the derivation group.The ECVm was larger in T2DM patients than that in HC?29.1±3.1%vs.26.4±2.4%,p=0.002?,whereas,the ECVsyn did not differ between T2DM and HC?28.4±3.0%vs.26.8±2.3%,p=0.051?.Compared with HC,HCM patients were associated with higher ECVsyn and ECVm of the mid-ventricle.6)Comparison of differences in ECVm and ECVsyn among HC,T2DM and HCM in the validation group.The ECVm was larger in T2DM patients than that in HC?28.6±2.9%vs.25.8±3.2%,p=0.002?,whereas,the ECVsyn did not differ between T2DM and HC?27.9±2.5%vs.26.7±2.7%,p=0.087?.Compared with HC,HCM patients were associated with higher ECVsyn and ECVm of the mid-ventricle.The ECVm of free wall was larger in HCM patients than that in HC,whereas,the ECVsyn of free wall did not differ between HCM and HC.7)Use of ECVsyn to categorize patients with abnormal ECV.Using our center's normal cut-off of 31.8%,the use of ECVsyn would lead to a 10–15%incorrect categorization of patients in the derivation groups and 8-18%in the validation groups.2.T2DM related LV extracellular matrix expansion.There were significant differences among controls,short term T2DM and long term T2DM in ECV?controls,26.7±2.7%;short term T2DM,28.4±2.9%;long term T2DM,28.8±3.6%,p=0.018?,in ECMVi?controls,13.3±2.8 ml/m²;short term T2DM,15.4±2.9 ml/m²;long term T2DM,15.6±3.8 ml/m²;p=0.008?,but not in MMi?p=0.098?,CVi?p=0.342?,myocardial native?p=0.580?or post-contrast T1?p=0.600?.3.T2DM related LV structural alterations.1)LV myocardial structural alterations.There were significant differences among controls,short term T2DM and long term T2DM in EDVi?controls,64.3±9.8 ml/m²;short term T2DM,65.7±7.6 ml/m²;long term T2DM,59.1±8.5 ml/m²,p=0.013?and MVR?controls,0.83±0.17 g/ml;short term T2DM,0.87±0.13 g/ml;long term T2DM,0.98±0.23 g/ml,p=0.005?,but not in MMi?controls,52.2±9.2 g/m²;short term T2DM,56.9±9.1 g/m²;long term T2DM,56.6±10.4 g/m²,p=0.098?.2)LV endocardial remodeling.There were significant differences among controls,short term T2DM and long term T2DM in max apical FD?controls,1.26±0.05;short term T2DM,1.28±0.04;long term T2DM,1.29±0.06,p=0.019?.3)Relationship between LV extracellular matrix expansion and structural alterations.ECMVi was associated with MVR?R=0.606,p<0.001?and max apical FD?R=0.279,p=0.008?,but not with EDVi?R=0.145,p=0.176?.Max apical FD was associated with EDVi?R=-0.426,p<0.001?and MVR?R=0.535,p<0.001?in all participants.4.T2DM related LV and LA functional alterations.1)LV functional alterations.There were significant differences among controls,short term T2DM and long term T2DM in longitudinal peak diastolic strain rate?PDSRL,controls,1.18±0.31 1/s;short term T2DM,1.07±0.17 1/s;long term T2DM,1.03±0.231/s,p=0.042?,but not in EF?p=0.878?or CI?p=0.164?.2)LA functional alterations.There were significant differences among controls,short term T2DM and long term T2DM in LA total EF?controls,57.5±8.6%;short term T2DM,56.9±7.4%;long term T2DM,49.5±9.2%,p<0.001?and passive EF?controls,36.5±10.7%;short term T2DM,32.3±9.2%;long term T2DM,28.2±8.8%,p=0.003?,but not in active EF?controls,32.3±12.5%;short term T2DM,36.3±7.0%;long term T2DM,29.5±11.0%,p=0.108?.3)Relationship between LV and LA functional alterations.PDSRL was associated with LA total EF?R=0.212,p=0.046?and active EF?R=0.212,p=0.046?.4)Relationship between LV structural alterations and LV and LA functional alterations.MVR was associated with LA passive EF?R=-0.284,P=0.007?and total EF?R=-0.240,P=0.023?.Max apical FD was associated with PDSRL?R=-0.265,P=0.012?5)CMR indices to discriminate among controls,short term T2DM and long term T2DM.Uni-and multi-variate logistic regression analysis showed that compared with controls,short term T2DM was only associated with ECMVi?odds ratio=1.30;95%CI:1.04¹.62;p=0.019?.The area under the receiver operating characteristic?ROC?curve?AUC?was 0.704.Uni-and multi-variate logistic regression analysis showed that compared with short term T2DM,long term T2DM was associated with EDVi?odds ratio=0.90;95%CI:0.81⁰.99,p=0.028?and LA total EF?odds ratio=0.90;95%CI:0.83⁰.98,p=0.021?.The AUC values for EDVi,LA total EF and combination of the two indices were 0.713,0.737 and 0.792,respectively.Conclusion1.The accuracy T1_blood-based HCTsyn in 3.0 T CMR.T1_blood-based HCTsyn and ECVsyn may lead to incorrect conclusions for populations with a subtly elevated ECV in clinical research,such as in T2DM,and can lead to significant miscategorization for individual patients in clinical care.So,in the following parts of this study,HCTm,not HCTsyn,was used to calculate ECV.2.T2DM related LV extracellular matrix expansion.In the early stage,T2DM patients may be associated with extracellular matrix expansion.ECMVi may be more sensitive index than ECV to detect T2DM related LV extracellular matrix expansion.3.T2DM related LV structural alterations.In the late stage,T2DM patients may be associated with endocardial remodeling and LV myocardial concentric remodeling,but not concentric hypertrophy.There may be significant relationship between LV extracellular matrix expansion and structural alterations.4.T2DM related LV and LA functional alterations.In the late stage,T2DM patients may be associated with LV subtle diastolic dysfunction and decreased LA function.There may be significant relationship among LV structural alterations and decreased LV and LA function.ECMVi may be the most sensitive index of T2DM related early alteration.EDVi and LA total EF may be the most sensitive indices of T2DM related late alteration.