Estimation Of Personal Exposure To Tobacco Smoke And Its Association With Abnormal Bone Mineral Density Based On Urinary Concentrations Of Nicotine And Its Metabolites

Part One: Evaluation of personal exposure to tobacco smoke in the elderly based on urinary concentrations of nicotine and its metabolitesObjective: The current study aimed to evaluate the advantages and disadvantages of urinary concentrations of nicotine and its metabolites in assessing active and passive exposure to tobacco smoke among community-dwelling elders as well as to provide scientific clues for quantifying personal tobacco smoke exposure and more accurate estimating risk of smoking-related diseases.Methods: Participants(n=6197)in this study were recruited from the baseline survey(n=9411)of the Shenzhen ageing-related disorder cohort,after excluding individuals with missing data for urinary concentrations of nicotine and its metabolites,creatinine,exposure to both active and passive smoking as well as patients with renal dysfunction and proteinuria.We measured urinary concentrations of nicotine(Nic)and its ten metabolites(including cotinine(Cot),cotinine N-β-D-glucuronide(Cot Gluc),(S)-nicotine-N-β-glucuronide(Nic Gluc),trans-3’-hydroxy cotinine(OHCot),trans-3’-hydroxy cotinine O-β-D-glucuronide(OHCot Gluc),(S)-cotinine N-oxide(CNO),rac 4-hydroxy-4-(3-pyridyl)butanoic acid dicyclohexylamine salt(Hy Py But),(1’S,2’S)-nicotine-1’-oxide(NNO),(R,S)-norcotinine(Nor Cot),(R,S)-nornicotine(Nor Nic))by using high-performance liquid chromatography-tandem mass spectrometry,in addition to urinary creatinine concentrations by a fully automated biochemical analyzer.We conducted receiver operator characteristic curve(ROC)of each study variable to assess active and passive exposure to tobacco smoke based on urinary concentrations of Nic or each of its metabolites(Nic Gluc,Cot,Cot Gluc,OHCot,OHCot Gluc,NNO,CNO,Hy Py But,Nor Nic or Nor Cot)or ∑Nic(Nic plus Nic Gluc),∑Cot(Cot plus Cot Gluc),∑OHCot(OHCot plus OHCot Gluc),total nicotine equivalent(TNE,including TNE2(∑Cot plus ∑OHCot),TNE3(TNE2 plus ∑Nic),TNE7(the sum of TNE3,urinary NNO,CNO,Nor Nic and Nor Cot concentrations)or TNE8(TNE7 plus Hy Py But)).Thereafter,we calculated the area under the curve(AUC)as well as the corresponding optimal cutoff point and its Youden index,specificity,sensitivity,false positive rate,false negative rate of the ROC curve of each study variable to evaluate the diagnosis value of these study variables on assessing personal tobacco smoke exposure.Results: In all participants: i)the estimated methods for identifying current smokers was better(all AUCs>0.85)based on urinary concentrations of Nic or each of its metabolites(Nic Gluc,Cot,Cot Gluc,OHCot,OHCot Gluc,NNO,CNO,Hy Py But or Nor Cot)or ∑Nic,∑Cot,∑OHCot,TNE2,TNE3,TNE7 or TNE8,except for urinary Nor Nic concentrations,of which these two methods(both Youden index equal to 0.86)for identifying current smokers based on urinary concentrations of Cot or CNO were better than the other methods based on urinary concentrations of Nic or each of other nicotine metabolites;ii)the estimated method for identifying current smokers based on urinary TNE2,TNE3,TNE7 or TNE8(all Youden index ranged from 0.87 to 0.88)was slightly better than those which were based on urinary concentration of Nic,each of nicotine metabolites(all Youden index<0.87),but the estimated methods based on urinary ∑Cot(Youden index=0.87)was more reasonable to identify current smokers rather than TNE2,TNE3,TNE7 or TNE8 considering the lower-cost detection of urinary ∑Cot.When estimations of active exposure to tobacco smoke were conducted in the NMR subgroups based on urinary concentrations of Nic,each of nicotine metabolites,∑Nic,∑Cot,∑OHCot,TNE2,TNE3,TNE7 or TNE8,the optimal cut-off points in the low NMR subgroup were higher than those in the high NMR subgroup,except for Cot Gluc,OHCot,OHCot Gluc,TNE3 or TNE7,indicating that there was some misclassification for distinguishing between current smokers and nonsmokers based on the same cut-off point in different NMR subgroups.In all participants,gender subgroups or NMR subgroups(all AUCs<0.7 and all Youden index <0.25),the estimation methods for passive exposure to tobacco smoke were unsatisfactory based on urinary concentrations of Nic,each of nicotine metabolites or ∑Nic,∑Cot,∑OHCot,TNE2,TNE3,TNE7 or TNE8.Conclusions: In all participants,the estimated method for identifying current smokers based on urinary Cot or CNO concentrations was better than the methods based on urinary concentrations of Nic or each of other nicotine metabolites;the estimated method for identifying current smokers based on urinary concentrations of TNE2,TNE3,TNE7 or TNE8 was slightly better than the method based on urinary concentrations of Nic,each of nicotine metabolites,∑Nic or ∑OHCot,but the methods based on urinary ∑Cot concentrations was more reasonable to identify current smokers rather than TNE2,TNE3,TNE7 or TNE8.In all participants,gender or NMR subgroups,the method was unsuitable to determine the current status of passive smoking based on urinary concentrations of Nic or each of nicotine metabolites,∑Nic,∑Cot,∑OHCot,TNE2,TNE3,TNE7 or TNE8.Part two: Association between exposure to tobacco smoke and abnormal bone mineral density: a mediating effect of relative telomere lengthObjective: The present study aimed to assess the associations between the current status of active or passive exposure to tobacco smoke by the three methods(based on selfreported tobacco smoke exposure,urinary Cot concentrations,self-reported tobacco smoke exposure with urinary Cot concentrations),urinary concentrations of nicotine metabolites and abnormal bone mineral density(a BMD,including osteopenia,osteoporosis),and analyze the mediating effects of leukocyte relative telomere length(LTL)on the associations.Methods: Based on participants(n=6197)involved in the first part of the study,we further excluded individuals with missing data for BMD or ones with family history of osteoporosis,self-reported patients with bone-metabolism related diseases,patients with drug therapy on bone metabolism and non-menopausal women,5841 individuals were finally included in this study.We measured LTL by real-time quantitative polymerase chain reaction.We conducted multinomial Logistic regression models to assess associations between personal exposure to tobacco smoke,urinary concentrations of eight nicotine metabolites or LTL and a BMD(osteopenia,osteoporosis)risk among gender subgroups,then we further determined nonlinear dose-response relationships between them above by using restrictive cubic spline functions(RCS).Thereafter,we employed quantile-based g computation approach to explore the joint effects of urinary concentrations of eight nicotine metabolites on a BMD risk.Finally,we analyzed the mediating effects of LTL on the correlations of urinary concentrations of nicotine metabolites with a BMD risk.Results: Multinomial Logistic regression models indicated that: in the males,after adjusted for the corresponding confounders and considering individuals with normal BMD as the reference group,the results showed that: i)self-reported current smokers were at higher risk of osteopenia than self-reported nonsmokers(odds ratio(OR): 1.33,95% confidence interval(CI): 1.05-1.69)or self-reported never smokers(OR: 1.36,95% CI: 1.06-1.74);ii)current smokers verified based on urinary Cot concentration were at higher risk of osteopenia than nonsmokers verified based on urinary Cot concentration(OR: 1.29,95% CI: 1.04-1.61);iii)current smokers verified based on self-reported active exposure to tobacco smoking with urinary Cot concentration were at higher risk of osteopenia than nonsmokers verified based on self-reported active exposure to tobacco smoke with urinary Cot concentration(OR: 1.37,95% CI: 1.10-1.71);iv)self-reported current smokers(OR: 1.69,95% CI: 1.19-2.41)or former smokers(OR: 1.71,95% CI: 1.23-2.37)were at higher risk of osteoporosis than self-reported never smokers.In the sex subgroups,after adjusted for the corresponding confounders and considering individuals with osteopenia as the reference group,multinomial Logistic regression models indicated that: i)only self-reported former smokers in males were at higher risk of osteoporosis than self-reported never smokers(OR: 1.62,95% CI: 1.21-2.16);ii)only nonsmokers with passive smoking verified based on urinary Cot concentration in females were at lower risk of osteoporosis than nonsmokers without passive exposure to tobacco smoke verified based on urinary Cot concentration(OR: 0.78,95% CI: 0.65-0.92).After adjusted for the corresponding confounders,a one-tertile increase in urinary concentrations(ln-transformed)of all 8 nicotine metabolites at the same time,quantilebased g computation approach indicated that: i)considering individuals with normal BMD as the reference group,males were at higher risk of osteopenia(OR: 1.27,95% CI: 1.04-1.55)or osteoporosis(OR: 1.50,95% CI: 1.14-1.97),but females were at lower risk of osteoporosis(OR: 0.63,95% CI: 0.43-0.94);ii)considering individuals with osteopenia as the reference group,only females were at lower risk of osteoporosis(OR: 0.62,95% CI: 0.50-0.76).Results from RCS suggested that,after adjusted for the corresponding confounders: i)considering individuals with normal BMD as the reference group,nonlinear doseresponse relationships were found between urinary concentration of Nic Gluc,OHCot or ∑OHCot and osteopenia risk in males,between only urinary Nic Gluc concentrations and osteoporosis risk in females(all the Poverall and Pnonlinear<0.05);ii)considering individuals with osteopenia as the reference group,nonlinear dose-response relationships were found between urinary concentrations of Nic Gluc,OHCot,NNO,∑OHCot,TNE2 or LTL and osteoporosis risk in females only(all Poverall and Pnonlinear<0.05).The mediation analysis suggested that in females: i)considering individuals with normal BMD as the reference group,LTL mediated the association of urinary Nic Gluc concentration with osteoporosis risk(the proportion of mediated effect: 9.37%);ii)considering individuals with osteopenia as the reference group,LTL mediated the association of urinary concentrations of Nic Gluc(the proportion of mediated effect:: 4.30%),Cot(the proportion of mediated effect: 8.31%)or OHCot(the proportion of mediated effect: 6.71%)with osteoporosis risk.Conclusions: The findings suggest that male current smokers had higher risk of a BMD(osteopenia,osteoporosis)rather than nonsmokers or never smokers,and there was nonlinear dose-response relationship between urinary Nic Gluc,OHCot or ∑OHCot concentrations and osteopenia risk in males;between urinary concentrations of Nic Gluc,OHCot,NNO,∑OHCot,TNE2 or LTL and with osteoporosis risk in females.Additionally,we found increased overall effect of urinary concentrations of each of eight nicotine metabolites on the a BMD(osteopenia,osteoporosis)risk in males and decreased overall effect on osteoporosis risk in females.LTL partial mediated the associations of urinary Nic Gluc,Cot,OHCot concentrations with osteoporosis risk in females.