Association Of Co-exposures To Metals And Nicotine With Obesity:the Mediating Effect Of Serum Uric Acid

Part One: Associations of urinary levels of metals or nicotine metabolites with obesity riskObjective: The study aimed to assess the associations of urinary concentrations of individual metal or individual nicotine metabolite with obesity risk.Methods: Participants in the study were from the baseline data(n=9411)of the Shenzhen Aging-Related Disorder Cohort in China.We excluded 2362 individuals who missed data regarding urinary concentrations of metals and nicotine metabolites,education levels,exercise habits,smoking status,drinking alcohol,diabetes,hypertension and body mass index,7049 ones were finally included in the analysis(body mass index(BMI)as a measure of obesity);subsequently,we continued to exclude 1912 individuals with missing data on waist circumference(WC),5137 ones were included to the second analysis(WC as another measure of obesity);thereafter,we did additional 6 ones who missed data on blood triglyceride,5131 ones were included to the third analysis(Chinese visceral adiposity index(CVAI)as a measure of obesity).Urinary concentrations of metals as well as nicotine and its metabolites were measured using inductively coupled plasma mass spectrometry and high-performance liquid chromatography with tandem mass spectrometry,respectively.Participants were then divided into three subgroups(<33.3%(T1),33.3-66.7%(T2)and ≥66.7%(T3),according to the corresponding tertiles of five obesity indices(including BMI,WC,a body shape index(ABSI),body roundness index(BRI)and CVAI,participants were divided into the subgroups: general(≥28kg/m2)and non-general obesity(BMI<28kg/m2)subgroups,central(WC≥90cm for male,≥85cm for female)and non-central obesity(<90cm for male,<85cm for female)subgroups,high(ABSI≥0.08)and low(<0.08)ABSI subgroups,high(≥4.0)and low(<4.0)BRI subgroups and high(≥110.5)and low(<110.5)CVAI subgroups.We assessed the associations of urinary individual metal or individual nicotine metabolite with risk for the five obesity types by binary Logistic regression models and non-linear dose-response relationships between above by using restricted cubic spline method,respectively.Results: Logistic regression analysis indicated that participants in the T2/T3 subgroup(the second or third tertiles,expressed as T2/T3)of urinary levels of titanium(1.27/1.32-fold),chromium(1.04/1.31-fold),vanadium(1.24/1.64-fold),zinc(1.30/1.39-fold),barium(1.29/1.48-fold),molybdenum(0.82/0.73-fold),cadmium(0.85/0.66-fold),cotinine N-β-D-glucuronide(Cot Gluc,1.09/1.35-fold),Rac 4-hydroxy-4-(3-pyridyl)butanoic acid dicyclohexylamine salt(Hy Py But,1.08/1.29-fold),(S)-cotinine N-oxide(CNO,1.15/1.40-fold)or(1’S,2’S)-nicotine-1’-oxide(NNO,1.11/1.48-fold)were at higher risk for general obesity than those in the T1 subgroup.Participants in the T2/T3 subgroup of urinary levels of titanium(1.28/1.33-fold),vanadium(1.30/1.40-fold),manganese(1.25/1.24-fold),cobalt(1.05/1.21-fold),barium(1.16/1.33-fold),thallium(1.42/1.35-fold),molybdenum(1.00/0.75-fold),cadmium(0.83/0.72-fold),Cot Gluc(1.04/1.26-fold),NNO(1.05/1.35-fold),CNO(1.05/1.33-fold)or NNO(1.09/1.24-fold)were at higher risk for central obesity than those in the T1 subgroup.Participants in the T2/T3 group of urinary levels of cobalt(1.14/1.35-fold),cadmium(1.21/1.37-fold)or Trans-3’-hydroxy cotinine(OHCot,0.81/0.76-fold)were at higher risk for high ABSI than those in the T1 subgroup.Participants in the T2/T3 group of urinary levels of titanium(1.17/1.25-fold),vanadium(1.32/1.31-fold),manganese(1.15/1.25-fold),cobalt(1.06/1.20-fold),zinc(1.20/1.29-fold),antimony(1.17/1.19-fold),barium(1.16/1.26-fold),thallium(1.37/1.54-fold),cadmium(0.86/0.80-fold),(S)-nicotine-N-β-glucuronide(Nic Gluc,1.30/1.22-fold),Cot Gluc(1.09/1.25-fold)or CNO(1.13/1.38-fold)were at higher risk for high BRI than those in the T1 subgroup.Participants in the T2/T3 group of urinary levels of vanadium(1.25/1.31-fold),manganese(1.27/1.26-fold),zinc(1.20/1.37-fold),tin(1.07/1.21-fold),barium(1.16/1.32-fold),thallium(1.26/1.32-fold),lead(1.13/1.23-fold),molybdenum(1.01/0.79-fold),cadmium(0.84/0.70-fold),Cot Gluc(1.30/1.22-fold),CNO(1.09/1.25-fold)or Cotinine(Cot,1.13/1.38-fold)were at higher risk for high CVAI than those in the T1 subgroup(all P-trend<0.05).Results from restricted cubic spline function showed that urinary levels of zinc,cadmium,antimony,Cot Gluc,Hy Py But,CNO or NNO had a non-linear relationship with risk for general obesity,urinary levels of zinc,molybdenum,Cot Gluc,CNO,Nic Gluc or NNO had a non-linear relationship with risk for central obesity,urinary levels of cadmium,CNO or Nic Gluc had a non-linear relationship with risk for high BRI;urinary levels of molybdenum,cadmium,Cot Gluc,CNO or Nic Gluc had a non-linear relationship with risk for high CVAI(all P for non-linear<0.05).Conclusions: We found the positive associations between urinary levels of titanium,chromium,vanadium,zinc,Cot Gluc,Hy Py But,CNO or NNO and risk for general obesity,between urinary levels of titanium,vanadium,manganese,cobalt,barium,thallium,Cot Gluc,Nic Gluc,CNO or Nic Gluc and risk for central obesity,between cobalt or cadmium and risk for high ABSI,between urinary levels of titanium,vanadium,manganese,cobalt,zinc,antimony,barium,thallium,Nic Gluc,Cot Gluc or CNO and risk for high BRI,between vanadium,manganese,zinc,tin,barium,thallium,lead,Cot Gluc,CNO or Cot and risk for high CVAI.We also found the negative associations between urinary levels of either molybdenum or cadmium and risk for general obesity,central obesity or high CVAI.Additionally,urinary levels of cadmium and OHCot Gluc were negatively associated with BRI and ABSI,respectively.Part Two: Associations of obesity with metal mixture,nicotine metabolites mixture or co-exposure to metal and nicotine metabolitesObjective: The study aimed to assess the associations of a mixture of either 22 metals or 8 nicotine metabolites in urine or environmental risk score(ERS)derived from co-exposure to metals and nicotine metabolites with obesity risk.Methods:Flowchart of participants included in Part Two was described in the Part One.We estimated the associations of a mixture of 22 metals or a mixture 8 nicotine metabolites in urine with five obesity types(general obesity,central obesity,high ABSI risk,high BRI risk and high CVAI risk)by quantile-based g-computation approach,and then selected the important predictors for obesity from the metals and nicotine metabolites as well as constructed environmental risk score(ERS)by elastic net models to evaluate the impacts of co-exposure to metals and nicotine metabolites.Participants were then divided into the three subgroups(<33.3%(T1),33.3-66.7%(T2)and ≥66.7%(T3)according to the ERS.Afterwards,we estimated the associations between ERS and obesity risk by binary Logistic regression models.Results: Results from quantile-based g-computation approach showed that with each tertile(33.3%)increase in urinary levels of a mixture of metals,the risk for general obesity,high ABSI and high BRI increased by 26%(OR: 1.26;95%CI: 1.03-1.54),23%(OR: 1.23;95%CI: 1.04-1.46)and 22%(OR: 1.22;95%CI: 1.05-1.42),respectively;furthermore,with each tertile(33.3%)increase in the urinary levels of a mixture of nicotine metabolites,the risk for general obesity,central obesity,high BRI and high ABSI increased by 35%(OR: 1.35;95%CI: 1.12-1.62),45%(OR: 1.45;95%CI: 1.21-1.75),36%(OR: 1.36;95%CI: 1.15-1.62)and 47%(OR: 1.36;95%CI: 1.15-1.62),respectively.The elastic net models selected three single pollutants(selenium,molybdenum,cadmium)and 74 pairwise interactions between metals and nicotine metabolites(molybdenum*molybdenum,cadmium*cadmium,vanadium*zinc,etc.).The ERS constructed from selected key variables ranged from ?3.08 to 3.99.Logistic regression models suggested the association of ERS with the risk for general obesity,central obesity,high BRI or high CVAI.In detail,individuals in the T2 or T3 subgroup(expressed as T2/T3)were at higher risk for general obesity(1.57/2.68-fold),central obesity(1.28/2.01-fold),high BRI(1.35/2.00-fold)or high CVAI(1.42/2.22-fold)compared to those in the T1 subgroup of ERS(all P<0.05).Conclusions: We found the positive associations between urinary levels of the mixture of metals and risk for general obesity,high ABSI and high BRI,between urinary levels of a mixture of nicotine metabolites and risk for general obesity,central obesity,high BRI and high ABSI.Co-exposure to metals and nicotine metabolites(represented by ERS)was positively associated with the risk for general obesity,central obesity,high BRI and high CVAI,but not for high ABSI.Part Three: The mediation effect of serum uric acid on the associations between co-exposure to metals with nicotine metabolites and obesityObjective: The study aimed to explore the mediation effects of serum uric acid(SUA)on the associations between co-exposure to metals with nicotine metabolites and obesity risk.Methods: Participants were from the baseline data(n=9411)of the Shenzhen Aging-Related Disorder Cohort in China.We then excluded 2377 individuals who missed data regarding urinary concentrations of metals,nicotine metabolites,education levels,exercise habits,smoking status,drinking alcohol,diabetes,hypertension,body mass index(BMI)and SUA,7034 ones were finally included in the analysis(BMI as a measure of obesity).Subsequently,we continued to exclude 1910 individuals with missing data on WC,5124 ones were included to the second analysis(WC as another measure of obesity);thereafter,we excluded one person who missed data on blood triglyceride,5123 ones were included to the third analysis(CVAI as a measure of obesity).Participants were then divided into the two subgroups according to the diagnostic criteria of hyperuricemia in Chinese males(SUA>420μmol/L)and females(SUA>360μmol/L).We then evaluated the mediation effect of SUA(as continuous or binary variable(hyperuricemia: yes/no))on the associations between ERS and obesity risk in the overall population and gender subgroups,respectively.Results: In the overall population,SUA(as continuous variable)partially mediated the association of ERS with general obesity or central obesity,the mediation effect accounted for 2.5% or 2.5% of the total effect of ERS on general obesity or central obesity;SUA(as binary variable)only partially mediated the association of ERS with general obesity(2.0%).In the females,SUA(as continuous or binary variable)also partially mediated the association of ERS with risk for general obesity(3.6%/2.4%),central obesity(5.6%/5.6%),high BRI(5.8%/5.0%)or high CVAI(8.1%/6.6%).No mediating effect of SUA on the associations of ERS and obesity risk was observed in the male.Conclusions: SUA partially mediated the associations of co-exposure to metals and nicotine metabolites with obesity risk in women but not in men.