The Study On The Association Of Renal Function Indicators With Cognitive Function And ATN Biomarkers And The Diagnostic Model Of Alzheimer’s Disease

With the rapid growth of aging population,cognitive impairment has become a worldwide problem that endangers human health.Alzheimer’s disease(AD)is the most common form of cognitive impairment,and there is no specific diagnostic methods and effective treatment currently.The amyloid deposition/tau neurofibrillary tangles/neurodegeneration/other related pathology(Amyloid/Tau/Neurodegeneration/Others,ATNX)framework provides a chance for andvancing AD diagnosis and treatment.Renal dysfunction is also a great challenge in senility,and the relationship between renal function and cognitive function has not been clarified.In this study,we analyzed the association of serum renal function indicators and cognitive function and Amyloid/Tau/Neurodegeneration(ATN)biomarkers in a Chinese cohort.We further explored if renal function indicators could be included into the ATNX framework to improve the diagnostic efficiency of AD.Part ⅠThe association of classic renal function indicators with cognitive function and ATN biomarkersObjective:To explore the association of renal function and cognitive function and ATN biomarkers and in different period and types of cognitive impairment.Methods:Participants from the Chongqing Ageing&Dementia Study(CADS)cohort who received systematic clinical neurological evaluation and complete data of serum renal function indicators from January 2012 to January 2021 were included,while neuropsychiatric disorders and other serious diseases of the system,tumors,special infections,etc.were excluded.Finally,a total of 427 participants were included.Classic renal function indicators include urea,creatinine,Chronic Kidney Disease Epidemiology Collaboration creatinine equation in estimating glomerular filtration rate(CKD-EPI eGFRCr),Chronic Kidney Disease Epidemiology Collaboration cystatin C equation in estimating glomerular filtration rate(CKD-EPI eGFRCysC),Chronic Kidney Disease Epidemiology Association glomerular filtration rate creatinine-Cystatin C estimation formula(CKD-EPI eGFRCr-CysC),Berlin Initiative Study equation in estimating glomerular filtration rate(BIS eGFR),Chinese modified Modification of Diet in Renal Disease Study equation in estimating glomerular filtration rate(cMDRD eGFR).The levels of renal function indicators among groups were compared using Student’s t-test,analysis of variance,Mann-Whitney U test or Kruskal-Wallis test.The chi-square test was used to compare the rates of categorical variables.Pearson or Spearman correlation analysis was used to assess the correlation between variables.Results:1.The association of classic renal function indicators and cognitive function:(1)Five types of eGFR levels were positively correlated with mini-mental state examination(MMSE)scores(R>0,p<0.001).After adjusting for covarietes,the positive correlation was still obvious between CKD-EPI eGFRCr-CysC levels and MMSE scores(p=0.002),but the correlation between CKD-EPI eGFRCysC levels and MMSE scores was negative(p=0.015).(2)According to the stage of cognitive impairment as the adjusted MMSE scores,the five eGFRs levels were positively correlated with the MMSE scores in the cognitive unimpairment(CU)stage(R>0,p<0.001).After adjusting for covariates,CKD-EPI eGFRCysC levels(p=0.005)and CKD-EPI eGFRCr-CysC levels(p=0.005)were still significantly positively correlated with MMSE scores.(3)With the increase of urea and creatinine levels and the decrease of eGFR levels,the proportion of mild cognitive impairment(MCI)and dementia increased.2.Dynamic changes of renal function indicators with the progression of cognitive impairment:CKD-EPI eGFRCysC levels were higher in moderate dementia than in early clinical stage(p=0.009),and CKD-EPI eGFRCr-CysC levels were lower in mild dementia than in old CU(p=0.001).Dynamic changes of renal function indicators with AD progression:serum urea levels were lower in AD with moderate dementia than in old CU(p=0.006),and CKD-EPI eGFRCr-CysC levels in AD with mild dementia were lower than in old CU(p=0.005).3.Differences of classic renal function indicators in different types of cognitive impairment and their correlation with cognitive function:there were no significant correlations between classic renal function indicators and cognitive function in different types of cognitive impairment;there was no significant difference in renal function indicators among different types of cognitive impairment.4.Correlations between classic renal function indicators and biofluid ATN biomarkers:the levels of CKD-EPI eGFRCysC were correlated with the cerebrospinal fluid(CSF)P-tau181(p=0.032),plasma Aβ42(p=0.013),plasma Aβ40(p=0.026)and plasma T-tau(p=0.002);the levels of CKD-EPI eGFRCr-CysC were correlated with plasma Aβ342(p<0.001)and plasma Aβ40(p<0.001);the levels of CKD-EPI eGFRCr-CysC were higher in the A+population than in the A-population(p=0.005);the levels of BIS eGFR were higher in the A+population than in the Apopulation(p<0.001),in the T+population than in the T-population(p=0.035),and in the N+population than in the N-population(p=0.006);the levels of cMDRD eGFR were higher in the A+population than in the A-population(p=0.009)and higher in the N+population than in the N-population(p=0.036).Conclusions:1.Renal dysfunction could be associated with cognitive impairment,and this association is obvious before the onset of clinical symptoms,suggesting that the influence of renal function may play a role in the early stages of cognitive impairment.There was no significant difference in classic renal function indicators among different types of cognitive impairment.2.With the progress of cognitive impairment,renal function gradually decreased and then slightly increased after reaching valley in the prodromal or mild dementia stage,suggesting that renal dysfunction mainly affects cognitive impairment at the early stage.3.Classic renal function indicators are correlated with ATN biomarkers,and decreased renal function could lead to aggravation of AD pathology.Part ⅡThe association of serum uric acid with cognitive function and ATN biomarkersObjective:Serum uric acid(sUA),as a powerful antioxidant,could play a neuroprotective role,but its superoxidative effect could also lead to nerve damage.Current studies focusing on the relationship between sUA and cognitive function are paradoxical.We aimed to investigate the association of sUA with cognitive function and ATN biomarkers in a clinical cohort that included different stages and types of cognitive impairment.Methods:Participants from the CADS cohort who received systematic clinical neurological evaluation and complete data of serum renal function indicators from January 2012 to January 2021 were included,while neuropsychiatric disorders and other serious diseases of the system,tumors,special infections,etc.were excluded.Finally,a total of 427 participants were included.The levels of sUA among groups were compared using Student’s t-test,analysis of variance,Mann-Whitney U test or Kruskal-Wallis test.The chi-square test was used to compare the rates of categorical variables.Pearson or Spearman correlation analysis was used to assess the correlation between variables.The best-fit model of the regression curve(linear,quadratic,cubic,or exponential)was screened by minimizing the Akaike information criterion(AIC)and maximizing the adjusted R2.Hierarchical regression analysis was used to assess moderating effects.Results:1.Age(p<0.001),gender(p<0.001),APOE gene(p=0.008)and CKD-EPI eGFRCr(p<0.001)were independent factors of sUA.2.The relationship between sUA and cognitive function:(1)The levels of sUA were positively correlated with MMSE scores(R=0.228,p<0.001),mainly in dementia stage(R=0.382,p<0.001).After adjusting for covarietes,the positive correlation was still obvious.(2)The proportion of MCI and dementia increased in the population with lower levels of sUA.3.The relationship between sUA and different types of cognitive impairment:(1)The correlation between sUA and cognitive function was mainly in AD(R=0.398,p=0.002)rather than non-Alzheimer’s disease cognitive dysfunction(non-AD).(2)The levels of sUA in Aβ+dementia were lower than Aβ-dementia(p=0.007).(3)Compared with CU and non-AD,sUA in AD had a more obvious downward trend with age.4.The relationship between sUA and ATN biomarkers:(1)The levels of sUA were positively correlated with CSF Aβ42,plasma Aβ42 and plasma Aβ40,and the correlation between plasma Aβ42 and sUA was still statistically significant after adjustment(p=0.019).(2)High levels of sUA reduced the correlation of CSF Aβ42 with MMSE scores(ΔR2=0.041,p<0.001)and the correlation of CSF P-tau with MMSE scores(ΔR2=0.024,p=0.001),which also alleviated the correlation of CSF Aβ42 with CSF P-tau181(ΔR2=0.037,p=0.001)and CSF T-tau(ΔR2=0.044,p<0.001).Conclusions:1.The levels of sUA were positively correlated with cognitive function especially in dementia and in AD.It means UA could be a protective factor of cognitive function especially in the dementia stage of AD.2.With the progression of AD,sUA levels showed a trend of increasing first and then decreasing,which reached peak at the prodromal stage of AD.3.The effect of sUA on cognitive function could be achieved through the interaction with AD pathological ATN biomarkers,and the neuroprotective effect of UA could act on confronting Aβ42 and the downstream pathological cascades.Part ⅢThe association of serum cystatin C with cognitive function and ATN biomarkersObjective:Serum cystatin C(sCysC)is an excellent indicator of renal function,and it has amyloid properties.To further analyze the association of sCysC with cognitive function and ATN biomarkers,especially Aβ,we conducted this part in the CADS cohort.Methods:Participants from the CADS cohort who received systematic clinical neurological evaluation and complete data of serum renal function indicators from January 2012 to January 2021 were included,while neuropsychiatric disorders and other serious diseases of the system,tumors,special infections,etc.were excluded.Finally,a total of 427 participants were included.sCysC levels among groups were compared using Student’s t-test,analysis of variance,Mann-Whitney U test or Kruskal-Wallis test.The chi-square test was used to compare the rates of categorical variables.Pearson or Spearman correlation analysis was used to assess the correlation between variables.Results:1.Age(p=0.002),CKD-EPI eGFRCr(p<0.001)and hypertension(p=0.018),were independent factors of sCysC.2.There is a negative correlation between sCysC and cognitive function(R=0.437,p<0.001)especially at the stage before the onset of clinical symptoms(R=0.339,p<0.001).(2)Subgroups with higher levels of sCysC(p<0.001)had higher portion of dementia and MCI.3.There were no significant differences of sCysC levels in different types of cognitive impairment.4.After adjusting for covariates,sCysC levels were lower in elderly CU than in mild dementia(p=0.029);sCysC levels were lower in old CU than in prodromal AD(p=0.035)5.The relationship between sCysC and ATN biomarkers:(1)The levels of sCysC in A+group were higher than in A-group(p<0.001).(2)sCysC levels were positively correlated with plasma Aβ42(p<0.001)and plasma Aβ40(p<0.001).Conclusions:1.The levels of sCysC are negatively correlated with cognitive function,and this correlation is prominent before the onset of clinical symptoms.2.The levels of sCysC gradually increase with the progress of AD and slightly decrease after prodromal AD.3.The levels of sCysC are related to CSF and plasma Aβ,and it suggests that CysC could participate in AD pathological process by interacting with Aβ.Part ⅣThe association of serum beta-2-microglobulin with cognitive function and ATN biomarkersObjective:β2-microglobulin(B2M)is an indicator of renal function,and it is rich in β-sheet structure and has amyloid properties.To further explore the association of B2M with cognitive function,AD and ATN biomarkers,we analyzed serum B2M(sB2M)in the CADS cohort.Methods:Participants from the CADS cohort who received systematic clinical neurological evaluation and complete data of serum renal function indicators from January 2012 to January 2021 were included,while neuropsychiatric disorders and other serious diseases of the system,tumors,special infections,etc.were excluded.Finally,a total of 427 participants were included.The levels of sB2M among groups were compared using Student’s t-test,analysis of variance,Mann-Whitney U test or Kruskal-Wallis test.The chi-square test was used to compare the rates of categorical variables.Pearson or Spearman correlation analysis was used to assess the correlation between variables.Results:1.Age(p<0.001),CKD-EPI eGFRCr(p<0.001),hypertension(p=0.007)and CHD(p=0.002)were independent influencing factors of sB2M.2.The levels of sB2M were negatively correlated with MMSE scores(R=-0.165,p=0.003),and the correlation was more obvious in the early stage(R=-0.306,p<0.001).The proportions of MCI and dementia in Q1 subgroup(with lower levels of sB2M)were lower than that in Q2(p=0.0018),Q3(p=0.009)and the Q4 subgroups(p<0.001).3.There were no significant differences of sB2M levels in different types of cognitive impairment.4.The levels of sB2M were positively correlated with plasma Aβ42(R=0.292,p<0.001)and plasma Aβ40(R=0.337,p<0.001)).Conclusions:1.The levels of sB2M are negatively correlated with cognitive function,and this correlation was more obvious before the onset of clinical symptoms.It is suggested that B2M affects cognitive function in the early stage of cognitive impairment,and may become one of the early diagnostic indicators of cognitive impairment.2.With the progress of cognitive impairment,the levels of sB2M first increased and then decreased,and reached the peak in the prodromal stage.3.The levels of sB2M are positively correlated with plasma Aβ42 and Aβ40,which means that B2M may be specifically related to AD and participate in the pathological process in the peripheral system..Part ⅤThe association of serum retinol binding protein with cognitive function and ATN biomarkersObjective:Retinol binding protein(RBP)is one of the indicators of renal function and is involved in the regulation of various metabolic pathways and physiological process.To further explore the association of RBP with cognitive function,AD and ATN biomarkers,we analyzed serum RBP(sRBP)in the CADS cohort.Methods:Participants from the CADS cohort who received systematic clinical neurological evaluation and complete data of serum renal function indicators from January 2012 to January 2021 were included,while neuropsychiatric disorders and other serious diseases of the system,tumors,special infections,etc.were excluded.Finally,a total of 427 participants were included.The levels of sRBP among groups were compared using Student’s t-test,analysis of variance,Mann-Whitney U test or Kruskal-Wallis test.The chi-square test was used to compare the rates of categorical variables.Pearson or Spearman correlation analysis was used to assess the correlation between variables.Results:1.Age(p=0.002),CKD-EPI eGFRCr(p<0.001)and gender(p<0.001)were independent influencing factors of sRBP.2.The levels of sRBP were negatively correlated with MMSE scores(R=-0.228,p<0.001).The proportions of MCI stage and dementia stage in the Q4 subgroup(with higher sRBP levels)were higher than that in the Q1 subgroup(p=0.011)and Q2 subgroup(p=0.038).3.sRBP had no significant difference between AD and non-AD.4.The levels of sRBP were positively correlated with CSF P-tau181(R=0.175,p=0.004)and CSF T-tau(R=0.184,p=0.002).Conclusions:1.The levels of sRBP are negatively correlated with cognitive function,and a high levels of sRBP are associated with a high proportion of MCI and dementia,which suggested that RBP could be a risk factor of cognitive function.2.sRBP levels gradually increase at first and then decreased with the progress of cognitive impairment,and the peak was in the period of MCI.3.sRBP levels are positive corelated to tau,which suggested that RBP could related to tauopathy.Part ⅥThe study for AD diagnostic model based on renal function indicatorsObjective:As mentioned above,renal function indicators are closely related to cognitive function and AD.This part continues to explore whether the above indicators can be included in the ATNX framework separately or combining with basic information,liver function,coagulation function,blood cell analysis,etc.and forming a comprehensive AD diagnostic model to improve AD diagnostic efficiency.Methods:Participants were included in the CADS cohort who received complete clinical neurological evaluation,renal function data and the data of CSF Aβ42 or Pittsburgh compound B positron emission tomography(PIB-PET)from January 2012 to January 2021,while neuropsychiatric disorders and other serious diseases of the system,tumors,special infections,etc.were excluded.General information,blood liver function,blood cell analysis,and coagulation assay data were collected.Student’s t-test and analysis of variance were used for the comparison between different groups of normally distributed continuous variables,and the Mann-Whitney U test and Kruskal-Wallis test were used for the comparison between different groups of non-normally distributed continuous variables.The chi-square test was used to compare the rates of categorical variables.logistic regression(forward method)is used for model fitting.The area under the curve(AUC)of the receiver operating characteristic(ROC)curve was used to assess the diagnostic performance of the model.The optimal model is selected according to the minimum AIC.The diagnostic cutoff value of the model can be derived from the Youden index(YI).Nomograph was used to easily visualize the contributions of the factors included in the model.Results:1.Single variable logistics regression analysis showed nine renal function indicators,twelve liver function indicators,six blood cell analysis indicators,three coagulation indicators and nine basic information indicators were included in the next analysis.2.The single renal function indicator with the highest diagnostic efficiency was CKD-EPI eGFRCr-CysC(AUC=0.740).3.The optimal renal function diagnostic model was Logit(P)=1.8664380.031744*BIS eGFR-0.004175*UA-0.019163*CKD-EPI eGFRCr-CysC+0.030712*RBP+0.029264*CKD-EPI eGFRCr(AUC=0.772).4.The optimal comprehensive model was Logit(P)=2.450283-0.121630*MMSE-0.008652*BIS eGFR+0.707020*b1+0.990589*APOE-0.048501*TP0.010392*CKD-EPI eGFRCysC(AUC=0.857).The maximum value of YI of the comprehensive model is 0.636,the corresponding cut-off value is-0.847,and the cutoff value can better distinguish AD and CU.Internal verification was carried out by 1000 repeated sampling correction experiments,and the results showed that the actual model fit the bias-corrected model well and was close to the ideal model(MAE=0.043).Conclusions:1.CKD-EPI eGFRCr-CysC has the highest diagnostic efficiency for AD in renal function indicators.2.The combined diagnosis of AD with renal function indicators,including BIS eGFR,UA,CKD-EPI eGFRCr-CysC,RBP and CKD-EPI eGFRCr,can improve the diagnostic efficiency.3.The optimal comprehensive model includes renal function(BIS eGFR,CKD-EPI eGFRCysC),liver function(b1,TP)and general information(MMSE,APOE).The diagnostic efficiency is higher than 85%,and the results of repeated sampling internal validation indicate the model is reliable.Conclusions1.Renal dysfunction could damage cognitive function in the early stage of cognitive impairment,and classic renal function indicators could be used in the early diagnosis of AD.Impaired renal function may aggravate the pathology of AD.2.Special renal function indicators are related to different ATN markers according to their respective characteristics,and act on different stages of cognitive impairment with different effects on cognitive function;it is expected to be incorporated into the peripheral ATNX framework for AD diagnosis;it is dynamic with the progress of cognitive impairment and can be considered for AD course monitoring and follow-up.3.Among the renal function biomarkers,CKD-EPI eGFRCr-CysC has the highest diagnostic efficiency for AD.The combined diagnosis of AD with renal function biomarkers can futher improve the diagnostic efficiency.The diagnostic efficiency of renal function combined with liver function and general information in the diagnosis of AD was higher than 85%.