Association Between Plasma Metals(Calcium,Magnesium,Copper)and First Stroke And Its Subtypes

Background and Objective:Epidemiological investigations have shown that stroke has become a leading cause of mortality and disability worldwide in China.Despite great efforts to identify traditional risk factors and interventions focusing on modifying unhealthy lifestyles,which reduces the incidence of stroke to some extent,there are still great residual risks.In addition to controlling traditional risk factors,there is an urgent need to identify novel modifiable risk factors to more accurately guide primary prevention of stroke.Recently,the importance of metal elements including calcium?Ca?,magnesium?Mg?and copper?Cu?in relation to risk of stroke has been increasingly recognized.However,the previous related research conclusions were inconsistent,and there were few related studies in Chinese population.Moreover,most related studies were limited to a single metal element,and the interaction between multiple metal elements had not been discussed.Therefore,this study mainly aimed to examine the associations of Ca,Mg,Cu and their combined effect with first stroke and its subtypes in Chinese hypertensive populations.Methods:We conducted a nested case control study from a multicenter,prospective,community-observed“China Hypertension Registry Study”.From January 2017 to December 2017,115337 patients with hypertension in Shandong Rongcheng were screened,and we collected their baseline general information and blood samples.Data were linked to stroke data through December 31st,2018.Patients with stroke data from the Chinese centers for disease control and prevention?CDC,2013-2018?who had complete records?physical exam,questionnaire,biological samples?according to identification number were selected as cases.The initial sample consisted of 1354 incident cases and 1354 matched controls.Then,this study matched stroke cases with an equal number of controls?patients without stroke?for age±1 years,sex,and study site.A total of 1255 first stroke cases and 1255 controls were included in the final analysis.Data on questionnaire included demographic characteristics?such as age,gender,education?,lifestyle?such as smoking,alcohol consumption?,history of diseases?hypertension,diabetes and dyslipidemia?and medication use?antihypertensive drugs,glucose-lowering drugs,lipoprotein-lowering drugs and antiplatelet drugs?.Physical examination included measurement of height,weight,waist circumference,hip circumference,blood pressure and pulse.Liver and kidney function,fasting blood glucose,fasting lipid and homocysteine at baseline were measured using automatic clinical analyzers?Beckman Coulter AU680?at the core laboratory.Baseline plasma Ca,Mg and Cu levels were measured by inductively coupled plasma mass spectrometry?ICP-MS?.Because the total plasma Ca level was affected by the albumin concentration,Ca?mg/dL?was corrected for based on the serum albumin?g/dL?values according to the Payne formula:albumin-corrected Ca?mg/dL?=Ca+0.8×?4.0-albumin?.A statistical analysis was performed using the Statistical Package for Social Science software 20.0?SPSS,IL,USA?,R 3.5.2 and Python.Conditional logistic regression analyses,smooth curve fitting?penalized spline method?and elastic net regression model were conducted to explore the association between multiple plasma metal elements and risk of first stroke and its subtypes.Receiver operating characteristic?ROC?analysis was also performed to compare model1?traditional risk factors?and model2?traditional risk factors+three metal elements?to predict the occurrence of first stroke and its subtypes.Results:1.A total of 2510 participants were included.The median follow-up time and interquartile range were 17.97?10.00-20.50?months.Overall,the mean?SD?age was 70.75?8.06?years and 49.48%were males.80.32%were hypertensive patients.Among the 1255 incident stroke cases,1079 were ischemic stroke,171 were hemorrhagic stroke and 5 no type stroke.The mean?SD?albumin-corrected Ca,Mg and Cu were 9.49?±0.63?mg/dL,19.88?±1.90?mg/L and 1.01?0.17??g/mL,respectively.According to the reference values of mental elements recommended by Mayo Clinic laboratories,1781?70.96%?participants'plasma Ca levels were in the normal range,only 384?15.30%?participants had hypocalcemian and 345?13.74%?participants had hypercalcemia;2253?89.76%?participants'plasma Mg levels were normal,only 130?5.18%?participants had hypomagnesemia,127?5.06%?participants had hypermagnesemia;2366?94.26%?participants had normal plasma Cu levels,only 113 participants?4.5%?had hypocopperemia,and 31 participants?1.24%?had high plasma Cu levels.2.Conditional logistic regression analyses were conducted to explore the association between single plasma metal element and risk of first stroke and its subtypes.After adjustment for important confounders such as age,gender,study sites,BMI,baseline SBP,DBP,smoking and drinking status,fasting blood glucose,blood lipids,history of disease,and medication status,?1?Per SD increment in plasma albumin-corrected Ca was not associated with increased risk of first stroke,first ischemic stroke and first hemorrhagic stroke[first stroke:OR?95%CI?=0.99?0.91-1.09?,P=0.858;first ischemic stroke:OR?95%CI?=0.99?0.90-1.09?,P=0.846;first hemorrhagic stroke:OR?95%CI?=0.93?0.69-1.26?,P=0.656].We also converted albumin-corrected Ca from a continuous variable to a categorical variable?tertiles?.Compared to participants in T2?<8.66 mg/dL?of albumin-corrected Ca,there was a significantly increased risk of first stroke[T1:OR?95%CI?=1.37?1.10-1.70?,P=0.005;T3:OR?95%CI?=1.30?1.04-1.62?,P=0.020;P_trend=0.626]and first ischemic stroke[T1:OR?95%CI?=1.41?1.11-1.79?,P=0.005;T3:OR?95%CI?=1.34?1.05-1.71?,P=0.017;P_trend=0.653]for participants in both T1?<8.66 mg/dL?and T3??9.14 mg/dL?.However,there was no significant associaion between albumin-corrected Ca and first hemorrhagic stroke.Smooth curve fitting further showed a U-shaped association between albumin-corrected Ca and risk of first stroke and first ischemic stroke.?2?Every SD increment in plasma Mg was not associated with increased risk of first stroke,first ischemic stroke and first hemorrhagic stroke[first stroke:OR?95%CI?=0.99?0.90-1.08?,P=0.796;first ischemic stroke:OR?95%CI?=1.00?0.91-1.11?,P=0.969;first hemorrhagic stroke:OR?95%CI?=1.04?0.79-1.37?,P=0.776].We also converted Mg from a continuous variable to a categorical variable?tertiles?.Compared to participants in T2?19.14-20.66 mg/L?of plasma Mg,there was a significantly increased risk of first stroke[T1:OR?95%CI?=1.37?1.10-1.70?,P=0.005;T3:OR?95%CI?=1.28?1.03-1.60?,P=0.028;P_trend=0.533]and first ischemic stroke[T1:OR?95%CI?=1.36?1.07-1.72?,P=0.011;T3:OR?95%CI?=1.31?1.03-1.67?,P=0.03;P_trend=0.689]for participants in both T1?<19.14 mg/L?and T3??20.66 mg/L?;However,there was no significant associaion between plasma Mg and first hemorrhagic stroke.Smooth curve fitting further showed a U-shaped association between plasma Mg and risk of first stroke and first ischemic stroke.?3?For per SD increment in plasma Cu,the risk of first stroke increased by 17%?OR=1.17,P<0.001?,the risk of first ischemic stroke increased by 17%?OR=1.17,P=0.001?.We also converted plasma Cu from a continuous variable to a categorical variable?quartiles?.Compared to participants in Q1?<0.90?g/mL?of the plasma Cu,the Q2-Q4 group significantly increased the risk of first stroke and first ischemic stroke,and there was a dose-dependent increase in risk of first stroke and first ischemic stroke with plasma Cu(first stroke:OR_Q2=1.02;OR_Q3=1.17;OR_Q4=1.46;P_trend=0.002;first ischemic stroke:OR_Q2=1.05;OR_Q3=1.19;OR_Q4=1.44;P_trend=0.005).P for trend in the all models was significant,suggesting the linear association between plasma Cu and first stroke.Although the risk of first hemorrhagic stroke in the Q2-Q4 group increased compared with the Q1 group,there did not reach statistical significance(P_trend>0.05).Further analyses using smooth curve fitting confirmed that the associations of plasma Cu and first stroke and first ischemic stroke were linear.3.Elastic net model selected 3 metals?Ca,Mg and Cu?for first stroke and its subtype using Python to establish predictive plasma metal scores in order to assess the association between multiple metal exposures and incident first stroke.For per SD increment in predictive plasma metal scores,the risk of first stroke increased by 17%?OR=1.17,P<0.001?,the risk of first ischemic stroke increased by 17%?OR=1.17,P=0.002?,the risk of first hemorrhagic stroke increased by 16%,but there did not reach statistical significance?P=0.279?.We also converted predictive plasma metal scores from a continuous variable to a categorical variable?tertiles?.Compared to participants in T1 of predictive plasma metal scores,T2 and T3 groups could significantly increase the risk of first stroke and first ischemic stroke,and the incidence risk increased with predictive plasma metal scores increasing(first stroke:OR_T2=1.34;OR_T3=1.40;P_trend=0.002;first ischemic stroke:OR_T2=1.29;OR_T3=1.39;P_trend=0.005).Regardless of the predicted plasma metal scores as a continuous or a categorical variable,no significant association was observed between predicted plasma metal scores and the risk of first hemorrhagic stroke?P>0.05?.According to the stratified analyses,regardless of subgroup,predictive plasma metal scores was positively associated with first stroke and first ischemic stroke.Moreover,associations of predictive plasma metal scores with first stroke and first ischemic stroke were stronger among current smokers and non-hypetrensives[first stroke,no smoking:OR=1.10,95%CI=1.00-1.21;smoking:OR=1.41,95%CI=1.16-1.73,P for interaction=0.011;without hypertension:OR=1.39,95%CI=1.14-1.71;with hypertension:OR=1.12,95%CI=1.02-1.23,P for interaction=0.047;first ischemic stroke,no smoking:OR=1.19,95%CI=0.98-1.21;smoking:OR=1.45,95%CI=1.16-1.81,P for interaction=0.009;without hypertension:OR=1.42,95%CI=1.14-1.76;with hypertension:OR=1.11,95%CI=1.00-1.23,P for interaction=0.043].4.The results of ROC showed that the addition of multiple metal exposures?model 2?significantly improved the area under the curve?AUC?beyond conventional model?model 1?from 0.669 to 0.656 for first stroke?P difference=0.003?and from 0.679 to0.667 for first ischemic stroke?P difference=0.007?.However,there was no significant difference between the two models in predicting first hemorrhagic stroke?0.684 vs.0.671,P>0.05?.These findings suggested that the addition of multiple metal exposures significantly improved the prediction of first stroke,especially first ischemic stroke.Conclusions:Among middle-aged and elderly Chinese community-based population,especially among hypertensives,both low and high plasma Ca and Mg levels were associated with increased risk of first stroke,especially ischemic stroke,supporting a U-shaped association of plasma Ca and Mg with first stroke and first ischemic stroke.The reference ranges of plasma Ca and Mg were 8.66-9.14 mg/dL and 19.14-20.66mg/L,respectively.plasma Cu levels,despite mostly within normal range,were significantly associated with higher risk of first stroke,especially ischemic stroke.Moreover,combine the three metal elements could improve the prediction of first stroke.The findings underscore the need to further research to establish optimal range of plasma mental elements in Chinese population.By doing so,it would provide more specific clinical and nutritional guidelines on optimal mental elements levels for the primary prevention of stroke.