Childhood Exposure to Parental Smoking and Midlife Cognitive Function: The Young Finns Study

Kokoteksti

(1)UEF//eRepository DSpace Rinnakkaistallenteet. https://erepo.uef.fi Terveystieteiden tiedekunta. 2020. Childhood Exposure to Parental Smoking and Midlife Cognitive Function: The Young Finns Study Rovio, Suvi Oxford University Press (OUP) Tieteelliset aikakauslehtiartikkelit © The Author(s) 2020 All rights reserved http://dx.doi.org/10.1093/aje/kwaa052 https://erepo.uef.fi/handle/123456789/23664 Downloaded from University of Eastern Finland's eRepository.

(2) 1. Childhood Exposure to Parental Smoking and Midlife Cognitive Function:. T. RI P. Suvi P. Rovio, Jukka Pihlman, Katja Pahkala, Markus Juonala, Costan G. Magnussen, Niina. SC. Pitkänen, Ari Ahola-Olli, Pia Salo, Mika Kähönen, Nina Hutri-Kähönen, Terho Lehtimäki, Eero Jokinen, Tomi Laitinen, Leena Taittonen, Päivi Tossavainen, Jorma SA. Viikari, and Olli T.. AN. U. Raitakari. M. Correspondence to Dr. Suvi P. Rovio, Research Centre of Applied and Preventive. ED. Cardiovascular Medicine, University of Turku, Kiinamyllynkatu 10, 20520 Turku, Finland,. O. RI. G. IN. AL. U. N. ED IT. email: suvi.rovio@utu.fi, tel: +358405665036. © The Author(s) 2020. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e‐mail: journals.permissions@oup.com.. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. The Young Finns Study.

(3) 2. Author affiliations: Research Centre of Applied and Preventive Cardiovascular Medicine,. T. Turku and Turku University Hospital, Finland (Suvi P. Rovio, Jukka Pihlman, Katja Pahkala,. RI P. Costan G. Magnussen, Niina Pitkänen, Pia Salo, Olli T. Raitakari); Paavo Nurmi Centre, Sports. SC. & Exercise Medicine Unit, Department of Physical Activity and Health, University of Turku,. Turku, Finland (Katja Pahkala); Department of Medicine, University of Turku and Division of. AN. U. Medicine, Turku University Hospital, Turku, Finland (Markus Juonala, Jorma SA. Viikari); Menzies Institute for Medical Research, University of Tasmania, Hobart, Australia (Costan G.. M. Magnussen); Institute for Molecular Medicine Finland (FIMM), University of Helsinki, Helsinki,. ED. Finland (Ari Ahola-Olli); Satakunta Central Hospital, Pori Finland (Ari Ahola-Olli); Department of Clinical Physiology, Tampere University Hospital and Faculty of Medicine and Health. ED IT. Technology, Tampere University, Tampere, Finland (Mika Kähönen); Department of Pediatrics, Tampere University Hospital and Faculty of Medicine and Health Technology, Tampere University, Tampere, Finland (Nina Hutri-Kähönen); Department of Clinical Chemistry, Fimlab. U. N. Laboratories and Finnish Cardiovascular Research Center-Tampere, Faculty of Medicine and. AL. Health Technology, Tampere University, Tampere, Finland (Terho Lehtimäki); Department of Paediatric Cardiology, Hospital for Children and Adolescents, University of Helsinki, Helsinki,. IN. Finland (Eero Jokinen); Department of Clinical Physiology, University of Eastern Finland and. G. Kuopio University Hospital, Kuopio, Finland (Tomi Laitinen); Vaasa Central Hospital, Vaasa,. RI. Finland (Leena Taittonen); Department of Pediatrics, University of Oulu, Oulu, Finland (Leena. O. Taittonen, Päivi Tossavainen); Department of Clinical Physiology and Nuclear Medicine, Turku University Hospital, Turku, Finland (Olli T. Raitakari).. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. University of Turku, Turku, Finland; Centre for Population Health Research, University of.

(4) 3. Funding: This work was supported by the Academy of Finland: grants 286284, 134309 (Eye),. T. Insurance Institution of Finland; Competitive State Research Financing of the Expert. RI P. Responsibility area of Kuopio, Tampere and Turku University Hospitals (grant X51001); Juho. SC. Vainio Foundation; Paavo Nurmi Foundation; Finnish Foundation for Cardiovascular Research;. Finnish Cultural Foundation; The Sigrid Juselius Foundation; Tampere Tuberculosis Foundation;. AN. U. Emil Aaltonen Foundation; Yrjö Jahnsson Foundation; Signe and Ane Gyllenberg Foundation; Diabetes Research Foundation of Finnish Diabetes Association; and EU Horizon 2020 (grant. M. 755320 for TAXINOMISIS); and European Research Council (grant 742927 for. ED. MULTIEPIGEN project); Tampere University Hospital Supporting Foundation. CGM is supported by National Heart Foundation of Australia Future Leader Fellowship (100849). KP is. ED IT. supported by Academy of Finland research fellowship (322112). Conflict of interest: none declared.. AL. U. N. Running head: Parental Smoking and Midlife Cognitive Function. ABSTRACT. IN. Acknowledging the lack of previous knowledge, we studied whether childhood/adolescent exposure to parental smoking associates with midlife cognitive function leveraging the data from. G. the Cardiovascular Risk in Young Finns Study. A population-based cohort of 3,596. RI. children/adolescents aged 3-18 years was followed-up between 1980-2011. Cognitive testing. O. was performed in 2011 on 2,026 participants aged 34-49 years using a computerized test. Measures of childhood/adolescent second-hand smoke exposure were parental self-reports of smoking and participants’ serum cotinine levels. Participants were classified into: 1)no exposure (non-smoking parents, cotinine<1.0ng/mL); 2)hygienic parental smoking (1-2 smoking parents,. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. 126925, 121584, 124282, 129378 (Salve), 117787 (Gendi), 41071 (Skidi) and 322098; the Social.

(5) 4. cotinine<1.0ng/mL); and 3)non-hygienic parental smoking (1-2 smoking parents, cotinine ≥1.0ng/mL). Analyses were adjusted for sex, age, family socio-economic status, polygenic risk. RI P. were at higher relative risk (RR) for poor (lowest quartile) midlife episodic memory and. T. cholesterol. Compared with non-exposed, participants exposed to non-hygienic parental smoking. associative learning (RR=1.38, 95%CI=1.08-1.75) and a weak association was found for short. SC. term and spatial working memory (RR=1.25, 95%CI=0.98-1.58). The associations for those. exposed to hygienic parental smoking were non-significant (episodic memory and associative. U. learning: RR=1.19, 95%CI=0.92-1.54; short term and spatial working memory: RR=1.10,. AN. 95%CI=0.85-1.33). Avoiding childhood/adolescence second-hand smoking exposure promotes. M. adulthood cognitive function.. ED. Key words: CANTAB, cognitive function, environmental tobacco smoking, parental smoking,. ED IT. second hand smoking. IN. AL. U. N. Abbreviations: CI: Confidence interval RR: Relative risk SBP: Systolic blood pressure SD: Standard deviation SE: Standard error YFS: The Cardiovascular Risk in Young Finns Study. G. Using data from the Cardiovascular Risk in Young Finns Study (YFS), we have shown that a. RI. high cumulative burden of cardiovascular risk factors, including smoking, elevated blood. O. pressure and high serum low-density lipoprotein cholesterol level in childhood/adolescence is associated with worse cognitive function in midlife (1). Adverse associations were particularly observed for episodic memory and associative learning. These data suggest that the associations between early life cardiovascular risk factors and midlife cognitive function not only reflect the. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. score for cognitive function, adolescence/adulthood smoking, blood pressure and serum total.

(6) 5. tracking of risk factor levels from childhood to adulthood, but that the risk factors potentially. T. between midlife smoking and higher risk of late-life cognitive impairment and dementia (2, 3).. RI P. Additionally to active smoking, adulthood second-hand smoking has been linked to deficits in. SC. executive function (4), impaired memory (5), increased dementia risk in non-smoking women (6), and declined memory in elderly women (7). Furthermore, cross-sectional and short-term. AN. U. follow-up studies have found childhood exposure to second-hand smoking to associate with worse cognitive performance, and lower intelligence and academic achievement in childhood (8,. M. 9). However, there are no previous studies focusing on the longitudinal association between. ED. childhood/adolescence exposure to second-hand smoking and midlife cognitive function.. ED IT. The biological links are tenable, with animal studies suggesting that second-hand smoking may negatively affect neurodevelopment and hippocampal receptor expression (10-12) and induce oxidative stress in the brain (13-15). Moreover, experiments in rodents have also indicated that. U. N. peri-adolescent exposure to nicotine can afflict indirect neuronal damage (16, 17) and has lasting. AL. impairing effect on cognitive functioning (11). It is therefore plausible that exposure to secondhand smoking in childhood could affect cognitive function and that this association might carry. IN. over to adulthood. Using data from the YFS that followed a large population-based sample of. G. individuals from childhood to adulthood, we hypothesized that childhood/adolescence exposure. RI. to parental smoking, especially those exposed to poor parental smoking hygiene (i.e. participants. O. whose parents smoked in the presence of their children), have worse cognitive function at midlife than those not exposed to parental smoking in childhood.. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. start to influence cognitive function in childhood. Other studies have found an association.

(7) 6. METHODS Study population. T. RI P. with medical schools and their rural surroundings in Finland. The first cross-sectional study of. the YFS including 3,596 randomly selected children and adolescents (boys and girls; aged 3, 6,. SC. 9, 12, 15 and 18 years) was performed in 1980. The cohort has been regularly followed-up at 3-9. U. year intervals. The latest follow-up study was carried out in 2011. Detailed information on the. AN. study population and protocol is reported in the Web Appendix 1 and elsewhere (18). The study. M. was conducted according to the guidelines of the Declaration of Helsinki, and the study protocol. ED IT. Procedures and measurements. ED. was approved by local ethics committees. All participants gave their informed consent.. Cognition. In 2011, cognitive function was assessed using a computerized cognitive testing battery (CANTAB®, Cambridge Cognition, Cambridge, UK) including five tests: 1) Motor. N. screening test used as a training and screening tool to indicate difficulties in test execution, 2). U. Paired associates learning test measured visual and episodic memory and visuospatial. AL. associative learning (hereafter episodic memory and associative learning), 3) Spatial working. IN. memory test measured short term and spatial working memory and problem solving (hereafter. G. short term working memory), 4) Rapid visual information processing test measured visual. RI. processing, recognition and sustained attention, and 5) Reaction time test measured reaction and. O. movement speed and attention.. Each test produced several variables. Principal component analysis was conducted for each test to identify components accounting for the majority of the variation within the dataset. After. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. The YFS is an ongoing longitudinal population-based study conducted in five university cities.

(8) 7. distribution analyses, the Motor screening test component was excluded from further analyses. RI P. subjects participated in the cognitive testing, and 1,798 had complete cognitive data. All. T. procedure resulting in four variables (mean 0, standard deviation (SD) 1). Altogether 2,026. SC. available data was used, and therefore, the number of participants varies between the cognitive domains. Detailed description of the cognitive tests, validation of the YFS cognitive data and. AN. U. principal component analysis is reported in the Web Appendix 1 and elsewhere (19).. M. Parental smoking, serum cotinine and parental smoking hygiene.. ED. Parents of the YFS participants self-reported their current smoking habits at baseline (1980) and the first follow-up (1983). Parents who indicated that they currently smoked either occasionally. ED IT. or daily were designated as current smokers. The biological validity of the queried parental smoking data has been studied previously by comparing the parental responses of current smoking with the offspring’s serum level of the major metabolite of nicotine namely cotinine. U. N. (20), that has been defined as an objective biomarker for smoking exposure (21). An increasing. AL. median serum cotinine level was found together with increasing number of smoking parent’s while the proportion of participants with zero cotinine level decreased. Finally, according to the. IN. information on current parental smoking, participants were classified into groups: 1) non-. RI. G. smoking parents, and 2) one or two smoking parents.. O. Childhood fasting serum samples were collected in 1980, stored at -20ºC and analyzed in 2014. Serum cotinine concentration was quantified using two methods. First, serum cotinine was extracted into dichloroethane by the method of Feyerabend and Russell (22) with the. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. due to ceiling effect. Other components were normalized using rank order normalization.

(9) 8. concentrated extract measured by means of gas-liquid chromatography, and with a quantitation. T. isotope labelled internal standard for cotinine were added to the 0.1mL serum sample. Sample. RI P. was extracted with dichloromethane and determined by liquid chromatography-tandem mass. SC. spectrometry with quantitation limit at 0.10ng/mL. Subsequently, the mean of both cotinine measurements was calculated for each participant. In case of missing information in either. U. measurement, the one obtained was used. For statistical analyses the subjects were divided into:. M. AN. 1) low (0-0.99ng/mL) and 2) elevated (≥1.00ng/mL) serum cotinine groups.. ED. Similarly to our previous study, a variable indicating parental smoking hygiene was generated for those participants with data on parental smoking and serum cotinine (N=1,976) (20). A 3-. ED IT. level categorical variable was constructed: 1) no parental smoking (non-smoking parents and a serum cotinine level <1.0ng/mL; N=869); 2) hygienic parental smoking (at least one smoking parent and a serum cotinine level <1.0ng/mL; N=567); 3) non-hygienic parental smoking (at. AL. U. N. least one smoking parent and/or a serum cotinine level ≥1.0ng/mL; N=540).. Covariates. Age was defined in full years at the end of the year 2011. The sum of household. IN. income at baseline was used as an indicator of family socio-economic status after categorization. G. into: 1) very low (<17,000euros/year), 2) low (17,000–27,000euros/year), 3) intermediate. RI. (27,001–37,000euros/year), and 4) high (>37,000euros/year) income groups. Subjects’ own. O. smoking status was queried in all follow-up studies among those aged 12 years and older. Adolescent smoking status was defined from baseline (1980) or the first follow-up (1983). Participants aged below 12 years were considered non-smokers. Adulthood smoking status was. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. limit at 0.16ng/mL. This method has been reported in detail previously (23). Second, 0.5ng.

(10) 9. queried and defined from the data in the latest follow-up (2011) or, in case of missing. RI P. activity and diet were assessed using standardized questionnaires similarly to our previous. T. were used for measuring systolic blood pressure (SBP) and serum total cholesterol (24). Physical. SC. studies (25, 26). Genotyping was performed for 2443 samples using custom build Illumina Human 670k BeadChip at Welcome Trust Sanger Institute. Genotypes were called using. AN. U. Illuminus clustering algorithm (PMID: 17846035). Genotype imputation was done using Beagle software (27) and The Sequencing Initiative Suomi (SISu) as reference data. A polygenic risk. M. score for cognitive function (hereafter polygenic risk score) was calculated using LDpred and. ED. used as a proxy for childhood cognitive ability. The reliability of the polygenic risk score was studied by correlating the risk score with participants’ childhood academic performance as well. ED IT. as education years in adulthood (r=0.17, P<0.0001 for both). Detailed description on the methods is presented in the Web Appendix 1.. U. N. Statistical analyses. The statistical analyses were first performed to the queried information on. AL. parental smoking habit, second to the measured serum cotinine level and finally combining these information into parental smoking hygiene variable to highlight the possible role of non-hygienic. IN. exposure to parental smoking on the offspring cognitive function. Linear regression analyses. G. were conducted to examine the associations between childhood/adolescence exposure to parental. RI. smoking and midlife cognitive function. All regression analyses were conducted as multivariable. O. models including first sex, age and family income (model 1), and second, including additionally polygenic risk score, and childhood/adolescence smoking status, SBP and serum total cholesterol (model 2). Finally, all analyses were further adjusted for adulthood smoking status, SBP and. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. information, in the two previous adulthood follow-up studies (2007 or 2001). Standard methods.

(11) 10. serum total cholesterol (model 3). Sensitivity analyses for elevated serum cotinine level were. RI P. childhood/adolescence, and second to participants with serum cotinine concentration below. T. restricting the population first to participants who reported being non-smokers in. SC. 3ng/ml. We used the cut point of <3ng/ mL to indicate passive smoke exposure in the sensitivity analyses among non-smoking participants, given that values above this level among children. AN. U. have been suggested to indicate active smoking (28). Additionally, sensitivity analyses were performed combining the queried information on parental ever and current smoking to define the. ED. M. exposure groups in relation to hygienic / non-hygienic parental smoking.. To further elucidate the association between childhood exposure to non-hygienic parental. ED IT. smoking and midlife cognitive function the risk ratios for low cognitive function (defined as the lowest 25th percentile in cognitive function) were calculated using Poisson regression modelling. The relative risks were calculated for the cognitive domains systematically indicating significant. U. N. associations for serum cotinine level and parental smoking hygiene. For these analyses, the subjects were divided into: 1) low (lowest quartile) and 2) high (three highest quartiles) cognitive. AL. function groups. Sensitivity analyses were conducted using 20th and 33th percentile as cut-off. IN. values for low cognitive function. Additionally, to overcome the possible bias due to loss-to-. G. follow-up we conducted multiple imputation for the missing values on cognitive function and. RI. adulthood covariates and performed sensitivity analyses for the main outcome parental smoking. O. hygiene using the imputed data. Finally, propensity score matching was performed matching the participants exposed to non-hygienic parental smoking with those not exposed in relation to age, sex, childhood family income, childhood/adulthood systolic blood pressure, serum total. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. performed using different cut-off values. Additionally, sensitivity analyses were conducted.

(12) 11. cholesterol, adulthood smoking. More detailed description of the multiple imputation and the. T. were restricted to the participants with no missing data on exposure, outcome, or the covariates.. RI P. All analyses were performed using SAS 9.4 (SAS Institute Inc. Cary, North Carolina, USA) and. SC. P<0.05 was used as the level of significance.. AN. U. RESULTS Characteristics. M. The characteristics of the study population and the numbers of participants in each separate. ED. cognitive test in the three groups if the main exposure group parental smoking hygiene are shown in Table 1. The representativeness of the study population participating in the cognitive. ED IT. testing was examined in our previous study by comparing the baseline differences between the participants and non-participants (1). Participants were more often women and older than nonparticipants. In addition, they originated from families with higher income and had somewhat. U. N. better academic performance during childhood than the non-participants. No differences were. AL. observed in parental smoking exposure, serum cotinine levels or any other characteristics (results not shown).. IN. Parental smoking. G. The analyses for queried parental smoking habits and cognitive function at midlife showed an. RI. inverse association between childhood parental smoking exposure and midlife episodic memory. O. and associative learning (Table 2, Paired associates learning test, model 1: β=-0.115, SE=0.05, P=0.023; model 2: β=-0.113, SE=0.05, P=0.024, model 3: β=-0.115, SE=0.05, P=0.022). No associations were found for other cognitive domains.. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. propensity score matching is presented in the Web Appendix 1. All multivariable model analyses.

(13) 12. RI P. inverse association between elevated cotinine levels and midlife episodic memory and. T. The age, sex and family income adjusted analyses for childhood serum cotinine levels showed an. SC. associative learning (Table 2; Paired associates learning test, model 1: β=-0.168SD, SE=0.06,. P=0.009). Adjustments for polygenic risk score, childhood/adolescence smoking status, SBP and. AN. U. serum total cholesterol did not alter the association (model 2: β=-0.158SD, SE=0.07, P=0.023). Similarly, the results remained unchanged after further adjustments for adulthood smoking, SBP. M. and serum total cholesterol (model 3: β=-0.150SD, SE=0.07, P=0.030). Additionally, a. ED. significant association was observed for short term working memory in the analyses adjusted for age, sex, family income, polygenic risk score, childhood/adolescence smoking, SBP and serum. ED IT. total cholesterol (Spatial working memory test, model 2: β=-0.134SD, SE=0.07, P=0.039). No significant associations were observed for other cognitive domains.. U. N. Parental smoking hygiene. AL. Worse midlife episodic memory and associative learning was found in the participants who had been exposed to non-hygienic parental smoking in childhood/adolescence compared to. IN. participants with non-smoking parents in the age, sex and family income adjusted analyses. G. (Table 3; Paired associates learning test, model 1: β=-0.195SD, SE=0.07, P=0.008). Further. RI. adjustments for polygenic risk score, childhood/adolescence smoking, SBP and serum total. O. cholesterol (model 2: β=-0.188SD, SE=0.07, P=0.011) and further for adulthood smoking, SBP and serum total cholesterol (model 3: β=-0.181SD, SE=0.07, P=0.015) did not modify the results. Additionally, there was a negative association between exposure to non-hygienic parental. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. Serum cotinine.

(14) 13. smoking in childhood/adolescence and midlife short term working memory (Spatial working. T. P=0.016; model 3: β=-0.158SD, SE=0.07, P=0.026). No significant associations were found for. SC. RI P. other cognitive domains.. Parental smoking hygiene and risk of low cognitive function. AN. U. The age, sex and family income adjusted analyses showed higher risk for low midlife episodic memory and associative learning for the participants with childhood exposure to non-hygienic. M. parental smoking compared to the participants with non-smoking parents (Table 4; Paired. ED. associates learning test, model 1: RR 1.37; 95%CI 1.09, 1.73; P=0.008). The association remained unchanged after further adjustments for polygenic risk score, childhood/adolescence. ED IT. smoking, SBP and serum total cholesterol (model 2: RR 1.39; 95%CI 1.09, 1.76, P=0.007) and additionally for adulthood smoking, SBP and serum total cholesterol (model 3: RR 1.38; 95%CI 1.08, 1.75; P=0.009). Additionally, a non-significant association between non-hygienic parental. U. N. smoking and short term working memory was observed (Spatial working memory test, model 1:. AL. RR 1.22; 95%CI 0.98, 1.53; P=0.077, model 2: RR 1.24; 95%CI 0.98, 1.57; P=0.072, model 3:. G. IN. RR 1.25; 95%CI 0.98, 1.58; P=0.068).. RI. Sensitivity analyses. O. Sensitivity analyses were conducted using several cut-off values for serum cotinine concentration (Web Table 1) and restricting the population to participants who 1) reported being non-smokers in childhood/adolescence and 2) had serum cotinine level ˂3 ng/ml were conducted. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. memory test, model 1: β=-0.143SD, SE=0.07, P=0.032; model 2: β=-0.170SD, SE=0.07,.

(15) 14. for serum cotinine levels (Web Table 2) and parental smoking hygiene (Web Table 3).. T. and current smoking in defining parental smoking hygiene groups (Web Table 4) , for the. RI P. association between parental smoking hygiene and the risk of low cognitive function (Web Table. SC. 5), and using 20th and 33th percentile as cut-off values for low cognitive function (Web Table 6).. Finally, the results from the analyses using multiple imputed data and propensity score matching. AN. U. approach are presented in the Web Tables 7 and 8. All sensitivity analyses indicated similar directions for the associations between childhood/adolescence exposure to parental smoking and. M. midlife cognitive function than the main analyses. However, some variation in the strength of. ED. association was observed. Detailed results are presented in the Web Appendix 1.. ED IT. Additional analyses. The possible modification introduced by sex, age, polygenic risk score, birthweight, childhood physical activity and diet as well as participant’s own adolescence and adulthood smoking for the. U. N. association of serum cotinine level or parental smoking hygiene on cognitive function was. AL. analyzed by adding a multiplicative interaction term for each possible modifier separately in the age, sex, and family income adjusted model (model 1). No statistically significant interactions. RI. G. IN. were found (data not shown).. O. DISCUSSION This is the first study to focus on the longitudinal association between childhood/adolescence exposure to parental smoking and adulthood/midlife cognitive function. The present study. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. Additionally, sensitivity analyses were performed combining the queried information on ever.

(16) 15. showed that exposure to parental smoking in childhood/adolescence associates systematically. T. memory in midlife. Additionally, exposure to non-hygienic parental smoking in. RI P. childhood/adolescence was found to associate with increased risk of low episodic memory and. SC. associative learning, and low short term working memory compared to subjects with non-. smoking parents. Importantly, all associations were independent of the participants own smoking. AN. U. in childhood/adolescence and in adulthood.. M. Previously, childhood second-hand smoke exposure has been linked to worse cognitive function,. ED. lower intelligence and poorer academic achievement (8, 9) but the study designs have been either cross-sectional or the follow-up times have not covered the full life-course from childhood into. ED IT. adulthood (9). Cognitive development continues until young adulthood and the slope of the development varies between individuals (29). In our study, cognitive testing was done in adulthood/midlife when the cognitive function may be assumed to have reached its full potential. U. N. (30). Additionally, we took the possible age related variation in cognitive function into account. AL. by introducing age as a potential confounder in all our statistical models.. IN. In the YFS population, we see a negative linear association between age and episodic memory. G. and associative learning (i.e. Paired associates learning test) that equals to -0.05 SD’s per year. RI. during the age window between ages 34 and 49, while the association of age for short term. O. working memory (i.e. Spatial working memory test) equals to -0.04 SD’s per year (19). Putting our present findings into clinical perspective by comparing the point estimate of parental smoking exposure to the point estimate of age, we notice that the association of non-hygienic. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. with worse episodic memory and associative learning as well as with short term working.

(17) 16. parental smoking exposure on episodic memory and associative learning corresponds to ~3.5. RI P. learning test localize mainly to medial temporal lobes, specifically to hippocampus and. T. years in cognitive aging (19). Furthermore, the neural networks related to the Paired associates. SC. parahippocampal gyrus (31), which have critical roles in learning and memory. In fact, a causal. link has been established specifically between learning and hippocampal neuro-/synaptogenesis. AN. U. (32), which is partly regulated by brain-derived neurotropic factor (33) – a neurotrophic factor that is an important component for memory consolidation and learning (34). A recent animal. M. study suggested that environmental tobacco smoke exposure during postnatal development. ED. decreases the hippocampal brain derived neurotrophic factor levels causing alterations in the development of the brain, which might lead to lasting interference on cognitive function (35). In. ED IT. animal models, exposure to tobacco smoke has been reported to decrease cell proliferation and the total number of surviving neuronal cells in dentate gyrus (i.e. a cortical region in the. AL. U. function (36).. N. hippocampus which is critical for learning and memory) and by that to plausibly affect cognitive. In addition to episodic memory and associative learning, we found associations between. IN. childhood/adolescence exposure to non-hygienic parental smoking and short-term working. G. memory (i.e. Spatial working memory test) – a cognitive domain which neural networks localize. RI. mainly to prefrontal cortex (37). A widespread network between brain regions is necessary to co-. O. operate optimal working memory capacity, but the exact contribution from different brain regions has remained uncertain. It is, however, known that even if prefrontal cortex is mainly responsible for executive functions including working memory, additionally frontal and parietal. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. years age association while the association for short term working memory corresponds to ~5.0.

(18) 17. cortices as well as temporal lobe/hippocampus (31) all play an important role in intact working. T. the frontal, parietal and temporal lobes (29). Therefore, the capacity of working memory. RI P. increases during childhood and early adulthood (39) making this cognitive domain susceptible. SC. for disturbance from childhood through adolescence to early adulthood. In line with our. observations, previous animal studies have shown that adolescent nicotine exposure induces age-. AN. U. specific changes in dendritic structure throughout the central nervous system (40).. M. Our results highlight the importance of avoiding children’s exposure to parental/passive. ED. smoking. There are several ways to limit exposure to environmental tobacco smoke in childhood, such as implementation of smoking bans. However, it may be speculated that smoking ban. ED IT. regulations could cause additional tobacco smoke exposure to children by increasing smoking in private places (e.g. at home). Therefore, we additionally need to expand parents’ knowledge on the consequences that their smoking may have on their children’s health, and importantly, that. U. N. these consequences may be long-lasting and carry over to adulthood. Even if non-smoking. AL. parents is the preferred option, promoting hygienic smoking can limit the children’s exposure if relinquishing smoking is not possible.. IN. There are some limitations in the present study. First, cognitive function was measured once, and. G. therefore, the role of childhood exposure to parental smoking on changes in midlife cognitive. RI. function was not studied. Second, with its short half-life (even ≤48 hours) cotinine may be. O. considered as an objective indicator of short-term exposure to tobacco smoking and to provide an outlook to short-term/acute exposure to parental smoking. Simultaneously, queried smoking data may be considered to reflect smoking habits during a longer time period. In the YFS cohort only. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. memory (38). During adolescence, the brain remains in its active state of maturation, particularly.

(19) 18. few parents (9.0% of the mothers and 7.3% of the fathers) stopped smoking during the. RI P. is thus likely that those parents who had exposed their children at the time of cotinine. T. reported being current smokers at least one parent had been a regular smoker also previously. It. SC. measurement had exposed their children also at other times. Third, we did not have data on. prenatal exposure to second-hand smoking, which has previously been linked to lower/altered. AN. U. childhood cognitive performance, cognitive deficits or delay in cognitive development (41, 42). Unfortunately, the lack of prenatal parental smoking data in the YFS hinders us from studying. M. the possible independent and different associations of pre- and postnatal smoking exposure for. ED. adulthood cognitive function Fourth, the lack of data on the used tobacco product, frequency of smoking, and/or number of daily cigarettes, the puffing volume, intensity or frequency as well as. ED IT. on pulmonary structures of the parents hinders us from studying the possible inter-individual differences in the nicotine absorption. Fifth, the possibility of residual or unmeasured confounding as well as misclassification of parental smoking due to self-reported information. U. N. remains possible. The methods used in this study to collect information on parental smoking have been previously validated in general population (43). Simultaneously, a good agreement. AL. between maternal report of partners’ smoking habits and partner’s self-reports (44) as well as an. IN. underestimation in relation to self-reported smoking status (45) have been reported previously.. G. Sixth, as in all longitudinal studies there is a possibility of differential loss to follow-up in our. RI. study. However, the baseline differences between the participants and non-participants have been. O. studied in our previous study (1). Participants were found to be more often women and older than non-participants. Additionally, they originated from families with higher income and had somewhat better academic performance during childhood than the non-participants. Importantly,. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. offspring’s childhood, and at the same time, in 98% of those families where both parents.

(20) 19. no other differences were observed. Finally, with respect to the establishment of causality,. T. significant associations between childhood/adolescence exposure to parental smoking and. SC. RI P. midlife learning and memory functions.. The main strength of the YFS is its longitudinal design, systematically repeated follow-up. U. studies beginning in childhood, and well-phenotyped cohort. Another important strength to be. AN. highlighted is the possibility to validate the participants’ exposure to parental smoking with the. ED. M. use of an objectively measured biomarker (i.e. serum cotinine).. Our study showed that exposure to parental smoking in childhood/adolescence is associated with. ED IT. worse midlife episodic memory and associative learning as well as short term working memory independent of own smoking habits in adolescence and adulthood. Our results underline the possibility that exposure to parental smoking in childhood may have long-lasting associations on. U. N. cognitive function that might carry over to adulthood and midlife. This further highlights the. AL. importance of the preventive actions against children’s smoking exposure in childhood and. G. IN. adolescence in order to promote adulthood cognitive function.. RI. Acknowledgements. O. Authors: Suvi P. Rovio, Jukka Pihlman, Katja Pahkala, Markus Juonala, Costan G. Magnussen, Niina Pitkänen, Ari Ahola-Olli, Pia Salo, Mika Kähönen, Nina Hutri-Kähönen, Terho Lehtimäki,. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. observational studies are prone to bias. Importantly, despite these limitations we could still find.

(21) 20. Eero Jokinen, Tomi Laitinen, Leena Taittonen, Päivi Tossavainen, Jorma SA. Viikari, and Olli T.. T. Author affiliations: Research Centre of Applied and Preventive Cardiovascular Medicine,. RI P. University of Turku, Turku, Finland; Centre for Population Health Research, University of. SC. Turku and Turku University Hospital, Finland (Suvi P. Rovio, Jukka Pihlman, Katja Pahkala,. Costan G. Magnussen, Niina Pitkänen, Pia Salo, Olli T. Raitakari); Paavo Nurmi Centre, Sports. AN. U. & Exercise Medicine Unit, Department of Physical Activity and Health, University of Turku, Turku, Finland (Katja Pahkala); Department of Medicine, University of Turku and Division of. M. Medicine, Turku University Hospital, Turku, Finland (Markus Juonala, Jorma SA. Viikari);. ED. Menzies Institute for Medical Research, University of Tasmania, Hobart, Australia (Costan G. Magnussen); Institute for Molecular Medicine Finland (FIMM), University of Helsinki, Helsinki,. ED IT. Finland (Ari Ahola-Olli); Satakunta Central Hospital, Pori Finland (Ari Ahola-Olli); Department of Clinical Physiology, Tampere University Hospital and Faculty of Medicine and Health Technology, Tampere University, Tampere, Finland (Mika Kähönen); Department of Pediatrics,. U. N. Tampere University Hospital and Faculty of Medicine and Health Technology, Tampere. AL. University, Tampere, Finland (Nina Hutri-Kähönen); Department of Clinical Chemistry, Fimlab Laboratories and Finnish Cardiovascular Research Center-Tampere, Faculty of Medicine and. IN. Health Technology, Tampere University, Tampere, Finland (Terho Lehtimäki); Department of. G. Paediatric Cardiology, Hospital for Children and Adolescents, University of Helsinki, Helsinki,. RI. Finland (Eero Jokinen); Department of Clinical Physiology, University of Eastern Finland and. O. Kuopio University Hospital, Kuopio, Finland (Tomi Laitinen); Vaasa Central Hospital, Vaasa, Finland (Leena Taittonen); Department of Pediatrics, University of Oulu, Oulu, Finland (Leena. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. Raitakari.

(22) 21. Taittonen, Päivi Tossavainen); Department of Clinical Physiology and Nuclear Medicine, Turku. RI P. 121584, 124282, 129378 (Salve), 117787 (Gendi), 41071 (Skidi) and 322098; the Social. T. This work was supported by the Academy of Finland: grants 286284, 134309 (Eye), 126925,. SC. Insurance Institution of Finland; Competitive State Research Financing of the Expert. Responsibility area of Kuopio, Tampere and Turku University Hospitals (grant X51001); Juho. AN. U. Vainio Foundation; Paavo Nurmi Foundation; Finnish Foundation for Cardiovascular Research; Finnish Cultural Foundation; The Sigrid Juselius Foundation; Tampere Tuberculosis Foundation;. M. Emil Aaltonen Foundation; Yrjö Jahnsson Foundation; Signe and Ane Gyllenberg Foundation;. ED. Diabetes Research Foundation of Finnish Diabetes Association; and EU Horizon 2020 (grant 755320 for TAXINOMISIS); and European Research Council (grant 742927 for. ED IT. MULTIEPIGEN project); Tampere University Hospital Supporting Foundation. CGM is supported by National Heart Foundation of Australia Future Leader Fellowship (100849). KP is supported by Academy of Finland research fellowship (322112).. U. N. Expert technical assistance in data management and statistical analyses by Johanna Ikonen, Irina. AL. Lisinen, and Noora Kartiosuo is gratefully acknowledged.. O. RI. G. IN. Conflict of interest: none declared.. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. University Hospital, Turku, Finland (Olli T. Raitakari)..

(23) 22. REFERENCES. RI P. and Midlife Cognitive Performance: The Young Finns Study. J Am Coll Cardiol. 2017;69(18):2279-2289.. Rusanen M, Rovio S, Ngandu T, et al. Midlife smoking, apolipoprotein e and risk of. SC. 2.. T. Rovio SP, Pahkala K, Nevalainen J, et al. Cardiovascular Risk Factors From Childhood. U. dementia and Alzheimer’s disease: A population-based cardiovascular risk factors, aging. 3.. AN. and dementia study. Dement Geriatr Cogn Disord. 2010;30(3):277-284.. Llewellyn DJ, Lang LA, Langa KM, Naughton F, Matthews FE. Exposure to secondhand. Heffernan TM, O'Neill TS. Everyday Prospective Memory and Executive Function. ED IT. 4.. ED. cotinine measurement. BMJ. 2009;338(7695).. M. smoke and cognitive impairment in non-smokers: National cross sectional study with. Deficits Associated with Exposure to Second-Hand Smoke. J Addict. 2013;2013(160486). 5.. Heffernan TM, O’Neill TS. Exposure to second-hand smoke damages everyday. Chen, Ruoling, Dongmei Zhang, Yang Chen B et al. Passive Smoking and Risk of. U. 6.. N. prospective memory. Addiction. 2013;108(2):420-426.. 3.. Pan X, Luo Y, Roberts AR. Secondhand Smoke and Women’s Cognitive Function in. IN. 7.. AL. Cognitive Impairment in Women Who Never Smoke. Arch Intern Med 2012;172(3):271-. G. China. 2018;187(5):911-918. Yang S, Decker A, Kramer MS. Exposure to parental smoking and child growth and. RI. 8.. O. development: A cohort study. BMC Pediatr. 2013;13(1).. 9.. Chen R, Clifford A, Lang L, Anstey KJ. Is exposure to secondhand smoke associated with cognitive parameters ofchildren and adolescents?-a systematic literature review. Ann. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. 1..

(24) 23. Epidemiol. 2013;23(10):652-661. Chen K, Nakauchi S, Su H, Tanimoto S, Sumikawa K. Early postnatal nicotine exposure. Counotte DS, Spijker S, Van De Burgwal LH, et al. Long-lasting cognitive deficits. SC. 11.. RI P. moleculare cells during adolescence in rats. Neuropharmacology. 2016;101:57-67.. T. disrupts the α2∗nicotinic acetylcholine receptor-mediated control of oriens-lacunosum. resulting from adolescent nicotine exposure in rats. Neuropsychopharmacology.. AN. 12.. U. 2009;34(2):299-306.. Slotkin TA, Skavicus S, Card J, Stadler A, Levin ED, Seidler FJ. Developmental. M. neurotoxicity of tobacco smoke directed toward cholinergic and serotonergic systems:. 13.. ED. More than just nicotine. Toxicol Sci. 2015;147(1):178-189.. Khanna A, Guo M, Mehra M, Royal W. Inflammation and oxidative stress induced by. 14.. ED IT. cigarette smoke in Lewis rat brains. J Neuroimmunol. 2013;254(1-2):69-75. Mazzone P, Tierney W, Hossain M, Puvenna V, Janigro D, Cucullo L. Pathophysiological. N. impact of cigarette smoke exposure on the cerebrovascular system with a focus on the. U. blood-brain barrier: Expanding the awareness of smoking toxicity in an underappreciated. 15.. AL. area - Review. Int J Environ Res Public Health. 2010;7(12):4111-4126. Slotkin TA, Skavicus S, Card J, Levin ED, Seidler FJ. Amelioration strategies fail to. IN. prevent tobacco smoke effects on neurodifferentiation: Nicotinic receptor blockade,. G. antioxidants, methyl donors. Toxicology. 2015;333:63-75. Ghosh D, Mishra MK, Das S, Kaushik DK, Basu A. Tobacco carcinogen induces. RI. 16.. O. microglial activation and subsequent neuronal damage. J Neurochem. 2009;110(3):10701081.. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. 10..

(25) 24. Bhagwat S V, Vijayasarathy C, Raza H, Mullick J, Avadhani NG. Preferential effects of. T. nicotine and 4-(N-methyl-N-nitrosamine)-1-(3-pyridyl)-1-butanone on mitochondrial. RI P. glutathione S-transferase A4-4 induction and increased oxidative stress in the rat brain.. Raitakari OT, Juonala M, Rönnemaa T, et al. Cohort profile: The cardiovascular risk in. 19.. AN. young Finns study. Int J Epidemiol. 2008;37(6):1220-1226.. U. 18.. SC. Biochem Pharmacol. 1998;56(7):831-839.. Rovio SP Nevalainen J, Juonala M, et al. Cognitive performance in young adulthood and. 20.. ED. Study. Neuropsychology. 2016;30(5):532-542.. M. midlife – relations with age, sex, and education The Cardiovascular Risk in Young Finns. West HW, Juonala M, Gall SL, et al. Exposure to parental smoking in childhood is. ED IT. associated with increased risk of carotid atherosclerotic plaque in adulthood: The Cardiovascular Risk in Young Finns Study. Circulation. 2015;131(14):1239-1246. 21.. Torres S, Merino C, Paton B, Correig X, Ramírez N. Biomarkers of Exposure to. U. N. Secondhand and Thirdhand Tobacco Smoke: Recent Advances and Future Perspectives.. 22.. AL. Int J Environ Res Public Health. 2018;15(12). Feyerabend C, Russell MA. A rapid gas-liquid chromatographic method for the. IN. determination of cotinine and nicotine in biological fluids. J Pharm Pharmacol.. G. 1990;42(6):450-452. Kallio K, Jokinen E, Hämäläinen M, et al. Impact of repeated lifestyle counselling in an. RI. 23.. O. atherosclerosis prevention trial on parental smoking and children’s exposure to tobacco smoke. Acta Paediatr. 2006;95(3):283-290.. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. 17..

(26) 25. 24.. Porkka K V, Raitakari OT, Leino A, et al. Trends in serum lipid levels during 1980-1992. Telama R, Yang X, Viikari J, Välimäki I, Wanne O, Raitakari O. Physical activity from. RI P. 25.. T. Epidemiol. 1997;146(1):64-77.. 26.. SC. childhood to adulthood: a 21-year tracking study. Am J Prev Med. 2005;28(3):267–73.. Laitinen TT, Pahkala K, Magnussen CG et al. Ideal Cardiovascular Health in Childhood. AN. U. and Cardiometabolic Outcomes in Adulthood Cardiovascular Risk in Young Finns Study. Circulation 2012;125:1971-1978.. Browning BL, Zhou Y, Browning SR. A One-Penny Imputed Genome from Next-. M. 27.. 28.. ED. Generation Reference Panels. Am J Hum Genet 2018;103(3):338-348 Benowitz NL, Bernert JT, Caraballo RS, Holiday DB, Wang J. Optimal serum cotinine. ED IT. levels for distinguishing cigarette smokers and nonsmokers within different racial/ethnic groups in the United States between 1999 and 2004. Am J Epidemiol. 2009;169:236–248. 29.. Arain M, Haque M, Johal L, et al. Maturation of the adolescent brain. Neuropsychiatr Dis. Harada CN, Love MCN, Triebel K. Normal Cognitive Aging. Clin Geriatr Med.. AL. 30.. U. N. Treat. 2013;9:449-461.. 2013;29(4):737-752. De Rover M, Pironti VA, McCabe JA, et al. Hippocampal dysfunction in patients with. IN. 31.. G. mild cognitive impairment: A functional neuroimaging study of a visuospatial paired. RI. associates learning task. Neuropsychologia. 2011;49(7):2060-2070.. O. 32.. Doetsch F, Hen R. Young and excitable: The function of new neurons in the adult mammalian brain - review. Curr Opin Neurobiol. 2005;15(1):121-128.. 33.. Wei Z, Liao J, Qi F, Meng Z, Pan S. Evidence for the contribution of BDNF-TrkB signal. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. in children and young adults. The Cardiovascular Risk in Young Finns Study. Am J.

(27) 26. strength in neurogenesis: An organotypic study. Neurosci Lett. 2015;606:48-52. Lee JLC, Everitt BJ, Thomas KL. Independent cellular processes for hippocampal. Torres LH, Garcia RCT, Blois AMM, et al. Exposure of neonatal mice to tobacco smoke. RI P. 35.. T. memory consolidation and reconsolidation. Science. 2004;304(5672):839-843.. SC. disturbs synaptic proteins and spatial learning and memory from late infancy to early adulthood. PLoS One. 2015;10(8):1-21.. U. Zhao C, Deng W, Gage FH. Mechanisms and Functional Implications of Adult. AN. 36.. Neurogenesis. Cell. 2008;132(4):645-660.. Spellman T, Rigotti M, Ahmari SE, Fusi S, Gogos JA, Gordon JA. Hippocampal-. M. 37.. 2015;522(7556):309-314.. Curtis CE. Prefrontal and parietal contributions to spatial working memory - Review.. ED IT. 38.. ED. prefrontal input supports spatial encoding in working memory. Nature.. Neuroscience. 2006;139(1):173-180. 39.. Klingberg T. Development of a superior frontal-intraparietal network for visuo-spatial. Smith RF, McDonald CG, Bergstrom HC, Ehlinger DG, Brielmaier JM. Adolescent. AL. 40.. U. N. working memory. Neuropsychologia. 2006;44(11):2171-2177.. nicotine induces persisting changes in development of neural connectivity. Neurosci. Cornelius MD, Ryan CM, Day NL, Goldschmidt L, Willford JA. Prenatal Tobacco Effects. G. 41.. IN. Biobehav Rev. 2015;55:432-443.. RI. on Neuropsychological Outcomes Among Preadolescents. J Dev Behav Pediatr.. O. 2001;22(4):217-225.. 42.. Trasti N, Vik T, Jacobsen G, Bakketeig LS. Smoking in pregnancy and children’s mental and motor development at age 1 and 5 years. Early Hum Dev. 1999;55:137-147.. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. 34..

(28) 27. 43.. Patrick DL, Cheadle A, Thompson DC, Diehr P, Koepsell T, Kinne S. The validity of self-. Brion MJ, Leary SD, Lawlor DA, Smith GD, Ness AR. Modifiable maternal. RI P. 44.. T. 1093.. maternal age, diet, and smoking. Pediatr Res 2008, 63(6):593–598.. U. Gorber SC, Schofield-Hurwitz S, Hardt J, Levasseur G, Tremblay M. The accuracy of. AN. 45.. SC. exposures and offspring blood pressure: a review of epidemiological studies of. self-reported smoking: A systematic review of the relationship between self-reported. O. RI. G. IN. AL. U. N. ED IT. ED. M. and cotinine-assessed smoking status. Nicotine Tob Res. 2009;11(1):12-24.. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. reported smoking: A review and meta-analysis. Am J Public Health. 1994;84(7):1086-.

(29) No. %. RI PT. Non-hygienic parental smoking. No.. %. No.. 190 247. 43.4 56.6. Mean (SD). 151 179 149 163. 23.5 27.9 23.2 25.4. 99 126 105 98. 23.1 29.4 24.5 22.9. 0.0 (0.01.0). U. 1 50. 0.2 7.6. AL. 1471d 1507. 88 97. 0.041. 12.5 (5.2) 43.5 (5.2). <0.001 <0.001 <0.001. 3.9 (1.00343.7). <0.001c. 30.2 30.8 19.9 19.1. 0.0 (0.01.0). 0.0 12.1. Mean (SD). 21.6 23.4. <0.001 <0.001. 0.11 (1.03). 0.03 (1.01). -0.19 (0.99). <0.001. 1495. 0.02 (1.01). 0.02 (0.98). -0.08 (1.00). 0.106. 1464. 0.07 (1.01). 0.01 (1.01). -0.09 (1.01). 0.016. O. RI. 1364. G. 0 53. 117 119 77 74. P-value. 47.6 52.4. 10.2 (4.5) 41.2 (4.5). N. 1504. 197 217). ED. 10.8 (4.9) 41.8 (4.9). ED IT. 1507 1507 1457. %. M. 270 41.2 386 58.8. AN. 1507. men women Age, years at baseline at cognitive testing Family income at baseline, euros/year <17000 17000-27000 27001-37000 >37000 SMOKING Childhood serum cotinine, ng/mlb Participants current daily smoking at baseline at cognitive testing COGNITIVE FUNCTION Episodic memory and associative learning Short term working memory Visual processing,. IN. Sex. Mean (SD). Hygienic parental smoking. SC. No. of No parental smoking participants. U. DEMOGRAPHICS. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. Table 1. Characteristics of The Cardiovascular Risk in Young Finns Study (1980-2011) Population According to The Parental Smoking Exposurea. 28.

(30) 113.0 (12.3) 118.2 (13.7). 111.8 (11.4) 118.2 (14.1). RI PT. 0.10 (1.00). -0.06 (1.01). 0.503. 5.3 (0.9) 5.1 (0.9). 114.7 (11.5). 0.053. 121.0 (14.4). 0.004. 5.1 (0.9) 5.2 (1.0). <0.001 0.355. ED IT. 5.3 (0.9) 5.2 (1.0). M. Serum total cholesterol, mmol/l at baseline 1507 at cognitive testing 1494 SD=standard deviation.. ED. at cognitive testing 1501. AN. U. SC. 0.00 (1.00). a. Values are means (standard deviations) for continuous variables and numbers (percentages) for categorical variables. Family income was. measured as a sum of participants’ parents’ income, transformed to correspond with the current value of money. Principal component analyses. N. was used to calculate components indicating episodic memory and associative learning (Paired associates learning test), short term working. U. memory (Spatial working memory test), visual processing, recognition and sustained attention (Rapid visual information processing test), and reaction and movement speed and attention (Reaction time test) based on CANTAB® (Cambridge Cognition, Cambridge, UK) cognitive test. AL. battery. P-values are from linear models for the normally distributed continuous variables and from trend test for categorical variables. Numbers are medians and interquartile ranges for the non-normally distributed serum cotinine value.. c. P-value for non-normally distributed serum cotinine value is from Kruskall-Wallis-test.. d. Smoking has been queried from participants beginning at age 12 years. Therefore, the number of participants with information on baseline. G. IN. b. O. RI. current smoking is restricted.. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. recognition and sustained attention Reaction and 1342 movement speed and attention COVARIATES Systolic blood pressure, mmHg at baseline 1502. 29.

(31) RI PT. ED IT. ED. M. AN. U. SC. Table 2. Associations Between Reported Parental Smoking (A) and Childhood/adolescence Serum Cotinine Level (B) and Midlife Cognitive Function in the Cardiovascular Risk in Young Finns Study (1980-2011)a. COGNITIVE DOMAINS Model 1 Model 2 Model 3 Number of β coefficients P-value β coefficients P-value β coefficients P-value participants (SE) (SE) (SE) A: QUERIED PARENTAL SMOKING Episodic memory and associative learning 1,481 0.115 (0.05) 0.023 -0.113 (0.05) 0.024 -0.115 (0.05) 0.022 Short term working memory 1,617 -0.073 (0.05) 0.132 -0.075 (0.05) 0.120 -0.065 (0.05) 0.176 Visual processing and sustained attention 1,588 -0.055 (0.05) 0.268 -0.049 (0.05) 0.310 -0.036 (0.05) 0.458 Reaction time 1,459 0.069 (0.05) 0.180 0.071 (0.05) 0.171 0.084 (0.05) 0.106 B: CHILDHOOD/ADOLESCENCE SERUM COTININE LEVEL Episodic memory and associative learning 1,200 -0.168 (0.06) 0.009 -0.158 (0.07) 0.023 -0.150 (0.07) 0.030 Short term working memory 1,310 -0.114 (0.06) 0.061 -0.134 (0.07) 0.039 -0.124 (0.07) 0.060 Visual processing and sustained attention 1,283 -0.013 (0.06) 0.839 0.009 (0.07) 0.888 0.027 (0.07) 0.686 Reaction time 1,179 -0.087 (0.06) 0.180 -0.131 (0.07) 0.060 -0.122 (0.07) 0.082 a Values are β coefficients (standard errors) and p-values from linear regression models for the participants with reported parental smoking (A) and elevated serum cotinine value (≥1.00 ng/mL) (B). The participants with no reported parental smoking (A) and low serum cotinine value. N. (values 0-0.99 ng/mL) (B) were used as the reference group. The analyses are conducted for four cognitive domains: episodic memory and. U. associative learning (Paired associates learning test), short term working memory (Spatial working memory test), visual processing, recognition. AL. and sustained attention (Rapid visual information processing test), and reaction and movement speed and attention (Reaction time test). SE=standard error. Model 1 is adjusted for age, sex and family income. Model 2 is adjusted for model 1 covariates plus the polygenic risk score. IN. for cognitive function, adolescent smoking, and childhood/adolescence systolic blood pressure and serum total cholesterol. Model 3 is adjusted. O. RI. G. for model 2 covariates plus adulthood smoking, systolic blood pressure and serum total cholesterol.. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. 30.

(32) RI PT. Table 3. Association of Parental Smoking Hygiene and Midlife Cognitive Function in the Cardiovascular Risk in Young Finns Study (19802011). a. SC. Pvalue. AN. Pvalue. MODEL 2 β coefficient (SE). Pvalue. -0.101 (0.07) 0.136 -0.181 (0.07) 0.015 0.012. -0.057 (0.06) 0.371 -0.143 (0.07) 0.032 0.033. -0.061 (0.06) 0.340 -0.170 (0.07) 0.016 0.017. -0.055 (0.06) 0.384 -0.158 (0.07) 0.026 0.029. -0.012 (0.07) 0.862 -0.023 (0.07) 0.735 0.733. -0.009 (0.07) 0.886 0.002 (0.07) 0.978 0.995. -0.003 (0.07) 0.969 0.019 (0.07) 0.791 0.818. 0.116 (0.07) 0.092 -0.040 (0.07) 0.568 0.790. 0.115 (0.07) 0.094 -0.081 (0.08) 0.286 0.540. 0.114 (0.07) 0.098 -0.069 (0.08) 0.362 0.641. M. -0.097 (0.07) 0.149 -0.188 (0.07) 0.011 0.009. ED. ED IT. 1,269. MODEL 3 β coefficient (SE). -0.090 (0.07) 0.191 -0.195 (0.07) 0.008 0.005 1,295. N. 1,165. AL. U. EPISODIC MEMORY AND ASSOCIATIVE LEARNING Hygienic parental smoking Non-hygienic parental smoking p-value for trend SHORT TERM WORKING MEMORY Hygienic parental smoking Non-hygienic parental smoking p-value for trend VISUAL PROCESSING AND SUSTAINED ATTENTION Hygienic parental smoking Non-hygienic parental smoking p-value for trend REACTION TIME Hygienic parental smoking Non-hygienic parental smoking p-value for trend. Number of participants 1,185. MODEL 1 β coefficient (SE). U. COGNITIVE DOMAINS. a. IN. Values are β estimates (standard errors) and p-values from linear models. Parental smoking hygiene was categorized in three categories: 1) no. G. parental smoking = subjects with non-smoking parents and a serum cotinine level 0-0.99 ng/mL (used as the reference category), 2) hygienic. RI. parental smoking = subjects with at least one smoking parent and with a serum cotinine level 0-0.99 ng/mL; 3) non-hygienic parental smoking =. O. subjects with at least one smoking parent and with a serum cotinine level ≥1.0 ng/mL and subjects with non-smoking parents but with a serum. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. 31.

(33) RI PT. cotinine level ≥1.0 ng/mL (exposure from own smoking or passively from other environmental source than parents). The analyses are conducted. SC. for four cognitive domains: episodic memory and associative learning (Paired associates learning test), short term working memory (Spatial working memory test), visual processing, recognition and sustained attention (Rapid visual information processing test), and reaction and. AN. U. movement speed and attention (Reaction time test). SE=Standard error. Model 1 is adjusted for age, sex and family income. Model 2 is adjusted for model 1 covariates plus the polygenic risk score for cognitive function, adolescent smoking, and childhood/adolescence systolic blood pressure and. O. RI. G. IN. AL. U. N. ED IT. ED. M. serum total cholesterol. Model 3 is adjusted for model 2 covariates plus adulthood smoking, systolic blood pressure and serum total cholesterol.. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. 32.

(34) RI PT. Table 4. Association Between Parental Smoking Hygiene and Relative Risk of Low Cognitive Function in the Cardiovascular Risk in Young Finns Study (1980-2011).a. SC. MODEL 2 RR 95% CI. Pvalue. MODEL 3 RR 95% CI. Pvalue. U. Pvalue. 1.20 1.37. 0.93, 1.55 1.09, 1.73. 0.156 0.008. 1.21 1.39. 0.93, 1.57 1.09, 1.76. 0.153 0.007. 1.19 1.38. 0.92, 1.54 1.08, 1.75. 0.190 0.009. 1.09 1.22. 0.85, 1.40 0.98, 1.53. 0.493 0.077. 0.85, 1.43 0.98, 1.57. 0.467 0.072. 1.10 1.25. 0.85, 1.43 0.98, 1.58. 0.464 0.068. 1.10 1.24. ED. EPISODIC MEMORY AND ASSOCIATIVE LEARNING Hygienic parental smoking Non-hygienic parental smoking SHORT TERM WORKING MEMORY 1,295 Hygienic parental smoking Non-hygienic parental smoking. MODEL 1 RR 95% CI. AN. Number of participants 1,185. M. COGNITIVE DOMAINS. a. ED IT. Values are relative risks (RR) (95% confidence intervals) and p-values from Poisson regression models for low cognitive function (lowest. quartile in cognitive function separately for each cognitive domain). Parental smoking hygiene is categorized into three categories: 1) no parental smoking = subjects with non-smoking parents and a serum cotinine level 0-0.99 ng/mL (ref), 2) hygienic parental smoking = subjects with at least one smoking parent and with a serum cotinine level 0-0.99 ng/mL; 3) non-hygienic parental smoking = subjects with at least one smoking. N. parent and with a serum cotinine level ≥1.0 ng/mL and subjects with non-smoking parents but with a serum cotinine level ≥1.0 ng/mL (exposure. U. from own smoking or passively from other environmental source than parents). The analyses are conducted for two cognitive domains: episodic. AL. memory and associative learning (Paired associates learning test) and short term working memory (Spatial working memory test). The subjects were divided into two groups according to their performance in each of the cognitive domains: 1) low (lowest quartile) and 2) high (three highest. IN. quartiles) cognitive function. RR=relative risk, CI=confidence interval. Model 1 is adjusted for age, sex and family income. Model 2 is adjusted for model 1 covariates plus the polygenic risk score for cognitive function, adolescent smoking, childhood/adolescence systolic blood pressure. O. RI. total cholesterol.. G. and serum total cholesterol. Model 3 is adjusted for model 2 covariates plus adulthood education, smoking, systolic blood pressure and serum. Downloaded from https://academic.oup.com/aje/advance-article-abstract/doi/10.1093/aje/kwaa052/5815309 by Joensuun Yliopisto Kirjasto user on 08 April 2020. 33.

(35)