Obesity and sex- and age-specific income – evidence from the HUNT study

Obesity and sex- and age-specific income – evidence from the HUNT study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Obesity and sex- and age-specific income – evidence from the HUNT study Christina Edwards, Johan H. Bjørngaard, Jonas Kinge This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-2275770/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Elevated body mass index (BMI) has been found to be associated with lower income, especially among women, and increasing evidence suggests that this association is causal. However, there is limited knowledge about the sex-specific effect of BMI on income at different ages. The relationship between BMI and income may change with age due to, for example, BMI-related morbidities or discrimination. The aim of this study was to investigate the sex-specific causal link between BMI and income at different ages. Methods The age-, and sex-specific effects were estimated using an instrumental variable approach with genetic variants as instruments (i.e., Mendelian randomisation) in an effort to deal with reverse causality and omitted variables that may bias the relationship between BMI and income. We also reduced measurement error by using registry-based income and clinically measured height and weight. Findings Elevated BMI led to a reduced likelihood of working, and lower income. For females, increased BMI led to lower income throughout, and particularly at the end of, work life. For males, increased BMI led to lower income from age 49 onwards. Health sciences/Signs and symptoms Health sciences/Risk factors Figures Figure 1 Figure 2 Introduction Past studies have found that increased body mass index (BMI) leads to lower income for females, and that male income seems less affected, or unaffected, by BMI ( 1 – 6 ). However, whether the effect of BMI on income differs by age is not known. Studying the effects of BMI on income throughout a work life can give new insights into how BMI influences income. Adolescent obesity has been found to lead to reduced earnings in adulthood (age 24–36), especially for females ( 6 ). If wages or the probability of income is affected early in adulthood, the effects may influence earnings throughout work life. On the other hand, one may expect that increased BMI is less likely to influence income in early adulthood, because the risk of BMI-related morbidities is smaller in younger adults ( 7 , 8 ) and time lived with obesity is shorter. As age increases, so does the likelihood of ill health due to elevated BMI, which may influence income through increased absenteeism, presenteeism, job opportunities, and/or labour market participation. A central methodological challenge has been to adequately address endogeneity biasing the relationship between BMI and income. BMI may lead to increased morbidity and mortality, which may reduce work ability and thereby wages. BMI could also influence income via, for example, employment, job promotion, choice of occupation, or job placements. On the other hand, income may influence accessibility to food and exercise options, which could influence weight ( 9 ). Second, the association between BMI and income may be confounded by a multitude of measurable and unmeasurable variables. For example, personality traits may influence income and likelihood of weight gain. Third, measurement error in studies using self-reported height, weight, and income, may affect the estimated effect of BMI on income ( 10 , 11 ). Several studies have tried to overcome these endogeneity problems by employing an IV approach using instruments such as: past BMI ( 12 ), area-level BMI ( 13 ), BMI of a biological relative ( 1 , 2 , 14 ), or genetic variants associated with BMI ( 15 , 16 ). These studies have also typically found that there is an association between BMI and income for females, but not for males. The main explanations proposed for this finding are: that increased BMI is associated with poorer health and may lead to lower work ability and thus wages, that individuals with elevated BMI are discriminated against on the labour market ( 17 – 20 ), and that there is uncertainty about whether employing workers with obesity will lead to increased costs and perhaps reduced productivity ( 21 , 22 ). In this paper, we combined data on objectively measured height and weight, BMI-related genetic variants, and sociodemographic variables from a longitudinal health survey with registry data on income from tax records to study the effect of BMI on one-year age-specific income at ages 20–70 using multivariable adjusted and IV analyses. Materials And Methods We used data from two rounds of the Trøndelag Health Studies (The HUNT studies). These population-based cohort studies were conducted in a geographical region in Central Norway between August 1995 and June 1997 (HUNT2) and between October 2006 and June 2008 (HUNT3). In these studies, the height and weight of participants was measured, and biological materials were collected for genotyping performed using Illumina HumanCoreExome arrays. Detailed information about The HUNT Studies is available from Åsvold, Langhammer ( 23 ). We linked the data from the HUNT studies, using a de-identified key, to national data available from Statistics Norway: The Norwegian Tax Administration Database and The Norwegian Population Registry. The Norwegian Tax administration database contains information about income, defined as the sum of personal and entrepreneurial income, from 1967 to 2016. The Norwegian National Registry contains information about whether individuals were alive and living in Norway during the years between 1986 and 2016. The annual income reported for each year from 1967 to 2016 was adjusted to 2016 price levels using the Norwegian Consumer Price Index. We then calculated one-year age-specific annual historical income for ages 20 to 70, using participant’s birth year. For instance, to calculate income at age 30 for a person born in 1975, we used income measured in 2005. For income measured between 1986 and 2016, participants were excluded at the year of death or not living in Norway. For income measured before 1986, all individuals were included. Income was converted from Norwegian Kroner (NOK) to 2016 Euros (€ 1 = NOK 9.29) ( 24 ). Regression analyses We estimated the one-year age-specific effect of BMI on: i) annual probability of working, ii) annual income, and iii) log of annual income, iv) z-score of annual income v) annual income among workers, and vi) z-score of annual income among workers, for ages 20–70. All analyses were stratified by sex, because the effect of BMI on labour market outcomes such as income differ by sex. First, we conducted instrumental variable (IV) analyses using two-stage least squares (2SLS) regressions using a weighted allele score as an instrument for BMI. These analyses were adjusted for birth year and study period. The alleles used in the score were selected based on 97 genetic variants which have been found to be associated with BMI in GWAS studies ( 25 ). All, except one (rs12016871) of these variants were available in our dataset, and for this variant we followed Brandkvist, Bjørngaard ( 26 ) and used the variant rs4771122 as a proxy. The weighted allele score was calculated by summing the weights (beta-coefficients reported by Locke and co-workers’) of the BMI-increasing alleles for each participant. For each one-year age group we first estimated the effect of the instrument (Z) on BMI (Eq. 1 ) $${BMI}_{i}={\pi }_{0}+ {\pi }_{1}{Z}_{i} + {\pi }_{2}{Sex}_{i}+ {\pi }_{3}{birthyear}_{i}+ {\pi }_{4}{study}_{i}+ {v}_{i}$$ 1 Then, using the predicted values from the first-stage (Eq. 1 ) we estimated the effect of BMI on each of the income-related outcomes (I (1-6) : I 1 = probability of working, I 2 = annual income, I 3 = log of income, I 4 = Income z-score, I 5 = income among workers, I 6 = z-score of income among workers) for each age-group (Eq. 2 ) $${\text{I}}_{({1-6)}_{i}}={\beta }_{0}+ {\beta }_{1}{\widehat{BMI}}_{i}+ {\beta }_{2}{Sex}_{i}+ {\beta }_{3}{birthyear}_{i}+ {\beta }_{4}{study}_{i}+ {u}_{i}$$ 2 The likelihood of working was estimated as the likelihood of having an income above zero. Log(income) was computed at each age as the log of income, where 0 income was set as 0.01. The income z-score was calculated as z = (x-µ)/σ, where µ and σ were calculated within each one-year age group for males and females. As expected, some participants had zero income despite being alive and living in the country. We therefore conducted a set of analyses including only participants that were working (i.e., with a positive income). Such analyses may be informative but should be interpreted with caution because we are conditioning on a possible consequence of the exposure ( 27 , 28 ). Next, each of the outcomes (I 1-6 ) were analysed using multivariable analyses adjusted for sex, birth year, study period when BMI measure was taken (HUNT2 or HUNT3), marital status (married, unmarried, divorced, widowed), smoking status (former smoker, sometimes smoker, daily smoker, and never smoker), educational level (primary school, secondary school, higher education), and urbanity (rural or urban). These covariates were selected because they have commonly been used in previous literature ( 29 ). Instrumental variable assumptions and non-linear associations The instrumental variable assumptions and non-linearity in the relationship between BMI and income has been explored in a previous paper assessing the sex-specific effects of BMI on income using a similar sample, but without age-specific estimates ( 16 ). For the IV analyses to provide consistent results, the IV-assumptions must be satisfied. There are three main assumptions. 1. The weighted allele score must be associated with income ( relevance assumption ). An F-statistic > 10 is typically considered satisfactory ( 30 , 31 ), but does not guarantee that any specific estimate will be unbiased or sufficiently well powered for the hypothesis of interest ( 32 ). 2. The weighted allele score should not be associated with any confounders (measured or unmeasured) ( independence assumption ). 3. The weighted allele score should only influence income via BMI ( exclusion restriction ). In the study assessing the non-age specific effect of BMI on income, the validity of the IV-assumptions was investigated using the weighted median, weighted modal, and MR-Egger approaches, as well as within-family analyses ( 16 ). The findings suggested that there was little evidence of the issues typically associated with breaches in the IV-assumptions in studies using genetic variants as IVs (i.e., there was little evidence of directional pleiotropy, population stratification, assortative mating and dynastic effects). Nevertheless, we cannot be certain that assumptions 2 and 3 are fulfilled. We report F-statistics for each analysis at each age, to test the relevance assumption. Previous studies have found a stable effect from this allele score in different ages ( 28 , 33 ). There was no evidence of nonlinearities, in males nor females, at normal to higher levels of BMI, though some evidence of nonlinearity at lower (underweight) levels of BMI was found for males. However, the statistical power may have been insufficient to detect an effect. Birth year specific effects As a sensitivity analysis, we estimated birth year-specific effects. The aim of these analyses was to try to uncover generational effects. Here we calculated the mean income for ten-year age groups: 20–29, 30–39, 40–49, 50–59, and 60–69. Then we estimated the sex-specific effects of BMI on mean ten-year income, for participants born in or before 1940, and individuals born after 1940. We applied the same methods as described above. We did not stratify into finer birth year groups to avoid losing statistical power. Results Descriptive information The average age of participants at the time that BMI was measured was 53 (SD 17.1, range: 19–99) for males, and 53 (SD 17.7, range:19–96) for females. The average BMI was 27.1 (SD 3.8, range:15–53) for males, and 26.9 (SD 4.9, range:12–56) for females. The majority of males were. classified as overweight, born between 1950 and 1959, married or cohabitating, had a secondary school education, lived in urban areas, and were never or former smokers (Table 1). The majority of females were: classified as normal weight or overweight, were born between 1950 and 1959, were married or cohabitating, had a secondary school education, lived in urban areas, and were never smokers (Table 1). Plots showing how these variables changed by birth year can be found in the Supplementary Information Appendix (SI Appendix) (Fig. S1, Fig. S2, and Fig. S3). Regression Results Results for males At the start of working life, males seemed to benefit slightly from elevated BMI (Fig. 1 ). From age 22 until 49, the effect of BMI on income was small, and the association was both positive and negative. From age 49 onwards, however elevated BMI was negatively associated with income. The results from the IV-models indicated that the probability of earning an income decreased with increasing BMI from age 22 onwards (except at ages: 24, 35 ,40–43 ,69, and 70). The multivariable adjusted results coincided with the results from the IV analyses, but at most ages the coefficients estimated from the IV analysis were larger and less precise (Fig. 1 , SI Appendix, Table S1). The effect of BMI on income followed a similar trend to the association between BMI and the probability of earning an income (Fig. 1 ). The results from the IV-analyses indicated that from age 20–25 the effect of BMI on income was positive (range: 56.4 to 232.9, mean = 128.7), from age 26–30 the effect estimates were negative (range: -63.6 to -189.1, mean = -113.9). Between age 30 and 48 the effect estimates varied between positive and negative (range: -177.5 to 134.9, mean =-6.2). From age 49 onwards, the effect estimate was negative except at ages 65 and 66 (range: -1295.9 to 93.5, mean = -533.0). The F-statistic was between 237 and 820, and declined with age (SI Appendix, Table S7). The findings from the multivariable adjusted analyses were similar, however, the effect estimates were smaller and more precise (Fig. 1 , SI Appendix, Table S2). When estimating the effect of BMI on log income (SI Appendix, Table S3, Fig. S4), the effects attenuated somewhat, but the age trend was similar to that of the effect of BMI on income. The mean z-score of income was − 0.008 (range: -0.044 to 0.014) in the IV analyses and was − 0.001 (range: -0.010 to 0.007) in the multivariable adjusted analyses (SI Appendix, Table S4). The income z-score and the conditional effect estimates (income given income and income z-score given income) followed a trend resembling that of the effect of BMI on income (SI Appendix, Table S5, Table S6, and Fig. S4). Results for females For females, the effects were more consistently negative across ages, with a tendency for increasing negative effects of BMI as age increased (Fig. 1 ). The results from the IV analyses indicated that there was a negative association between BMI and the likelihood of earning an income, except for between age 32 and 42 (in this age-interval the effect estimates were mainly positive) (Fig. 1 , SI Appendix, Table S8). The results from the multivariable adjusted analyses suggested that increased BMI was associated with a smaller likelihood of earning an income at all ages. The results from the IV-models indicated that the association between BMI and income was negative at all ages, except at ages 68–70, where the effect was positive (mean: -236.2, range: -606.2 to 46.7) (SI Appendix, Table S9). The F-statistic was between 212 and 812 and decreased with age (SI Appendix, Table S14). The results from the multivariable adjusted analyses were similar, with smaller effects and higher precision. In the multivariable adjusted analysis BMI was associated with a lower income at all ages (mean = -144.1, range = -290.7 to -16.6) When log income was used as the outcome measure, the effect estimates attenuated somewhat (SI Appendix, Table S10, Fig. S5), compared with when income was used. The mean z-score of income was − 0.007 (range: -0.013 to -0.002) in the multivariable adjusted analysis and − 0.012 (range: -0.027 to 0.007) in the 2SLS (SI Appendix, Table S11). The income z-score and the conditional effect estimates (income given income and income z-score given income) followed a similar trend to the effect of BMI on income (SI Appendix, Table S12, Table S13, and Fig. S5). Birth year-specific estimates To explore differences between age effects and generational effects, we estimated 10-year age-specific effects for males and females born before and after 1940. The estimates from the IV analysis suggest a slight tendency that the effect of BMI differs depending on the birth year category (Fig. 2 , bottom panel). For males there was no clear trend: the confidence intervals overlapped, and the effect estimates at different ages were both larger and smaller in the older and younger groups. For females, there was a clear tendency for that the effect of BMI on income was more negative for females born after 1940, compared with the younger group. The difference was most prominent for mean income between age 30 and 39. The multivariable adjusted analysis estimates were similar for males and females born before and after 1940 (Fig. 2 , top panel). Discussion We show that the relationship between BMI and income is negative throughout work life for females. For males, the effect of BMI on income was most prominent and negative at the end of working life (with the negative effects starting around age 49). No impact of BMI on income in males was found at younger ages. These findings were consistent across the income outcomes investigated in this study. Our study demonstrates that increased BMI also influences male income, but that these effects tend to appear closer to the end of work life and are less precise. There may be several explanations for this. BMI may influence the health of males and females differently. A previous study using national Norwegian data, found that healthcare use increased substantially for males around age 50 and onwards ( 34 ). A similar, but smaller increase was seen for females. This coincides with our study findings. In addition, there are other important differences between the sexes: males and females have different types of occupations, females are over twice as likely to work part time (16.7% of males and 37.3% of females work part time ( 35 ), females have maternity leaves, and are more likely to take time off to care for children, and the educational levels of males and females have changed over time. Moreover, the interplay between male and female wages within households may be relevant. The positive association between BMI and income at the youngest ages could be due to selection into higher education. If individuals with a higher BMI are more likely to select into jobs that do not require a higher education, then we would expect those with a higher BMI to have a higher income early in work life. When we stratified by birth year, we found that the effect of BMI on mean income from age 30–39 (and to some extent age 40–49) was positive for females born in 1940 or after, and negative for females born after 1940. This finding may have several explanations and should be explored further. First, labour market participation among females was lower before 1940 (SI Appendix, Fig. S6), and the most common occupations for females has also changed over time. In addition, ages 30–39 and 40–49 are ages that females are likely to be caring for children, and this may affect income. Rights and opportunities to work while caring for children has also changed considerably over the last decades. In Norway, the increase in population BMI seems to have begun in the 1960s for males and in the 1980s for females ( 36 ). Those born before 1940 may therefore have been less exposed to the environmental causes of obesity. The obesogenic environment may have affected males and females differently. Another explanation is that BMI-related labour market discrimination increased as more females entered the labour market. A strength of this study is that we have used registry-based income, and measured height and weight to compute BMI. However, using historical income to compute age-specific income, and the data informing BMI, genetic risk, and sociodemographic variables were measured either between 1995 and 1997 or between 2006 and 2008. This means that we are mixing cohorts and the time-of-measurement may bias the study findings. In addition, there may be a small degree of survival bias since individuals must have survived to participate in the study. The amount of bias from these factors may be different for males and females. We have tried to address this by conducting various sensitivity analyses, and by providing descriptive information at different ages and for different birth years (to the extent that his was possible) to improve transparency. Another limitation is that we have used the genetic variants presented by Locke, Kahali ( 25 ). Other more recent studies have revealed additional variants, but these were not available in our data. We also did not explore the age-specific extent of pleiotropy (i.e., that the BMI-related genetic variants influence other phenotypic traits that influence income). However, this has previously been done for the effect of BMI on male and female income ( 16 ). Our results, do suggest, however that the effect of BMI-related genetic variants on income increases with age (SI Appendix, Fig. S7 and Fig. S8). Lastly, we studied personal wages, rather than household income, and we did not consider how partners’ income could influence personal income. Our findings show that BMI influences income for both males and females and demonstrate the importance of conducting age-specific analyses. An increased understanding of the underlying mechanisms driving the age-specific effects for males and females identified in our study could provide clues about how to address income disparities driven by BMI and about how to reduce the societal costs of increased population BMI. Future studies, with larger sample sizes, should consider looking at age-specific effects for different occupational groups, and should also explore the role of height and education in the association between BMI and income. In addition, the effect of BMI on household income should be investigated as well as the effect of individual BMI on spouses’ income. Conclusion Increased BMI leads to lower income for both sexes. As BMI increases, females have a lower probability of earning an income and reduced annual income at most ages. For males BMI-driven reductions in income were evident from around age 49. Declarations Conflict of interest statement The authors have no conflicts of interest to declare. Funding The Norwegian Research Council; Grant number: 295989. Acknowledgements The Trøndelag Health Study (HUNT) is a collaboration between HUNT Research Centre (Faculty of Medicine and Health Sciences, Norwegian University of Science and Technology NTNU), Trøndelag County Council, Central Norway Regional Health Authority, and the Norwegian Institute of Public Health. Author Contributions All authors: conceptualization, reviewing and editing of manuscript, CHE: Formal analyses, and writing of original draft, JHB & JMK: Funding acquisition. Data availability statement The data that support the findings of this study are available from third parties. Restrictions apply to the availability of these data, which were used under license for this study. After de-identification, the individual participant data that underlie the results reported in this article are available to researchers whose proposed use of the data has been reviewed and approved by the data owners and Regional Committees for Medical and Health Research Ethics ( https://.portalen.no ) . References Brunello G, d’Hombres B. Does body weight affect wages?: Evidence from Europe. Economics & Human Biology. 2007;5(1):1–19. Cawley J. The impact of obesity on wages. Journal of Human resources. 2004;39(2):451–74. Han E, Norton EC, Stearns SC. Weight and wages: fat versus lean paychecks. Health economics. 2009;18(5):535–48. Baum CL, Ford WF. The wage effects of obesity: a longitudinal study. Health economics. 2004;13(9):885–99. Villar JG, Quintana-Domeque C. Income and body mass index in Europe. Economics & Human Biology. 2009;7(1):73–83. Amis JM, Hussey A, Okunade AA. Adolescent obesity, educational attainment and adult earnings. Applied Economics Letters. 2014;21(13):945–50. Thompson D, Edelsberg J, Colditz GA, Bird AP, Oster G. Lifetime health and economic consequences of obesity. Archives of internal medicine. 1999;159(18):2177–83. Khan SS, Ning H, Wilkins JT, Allen N, Carnethon M, Berry JD, et al. Association of body mass index with lifetime risk of cardiovascular disease and compression of morbidity. JAMA cardiology. 2018;3(4):280–7. Butland B, Jebb S, Kopelman P, McPherson K, Thomas S, Mardell J, et al. Tackling obesities: future choices-project report: Citeseer; 2007. Krul AJ, Daanen HA, Choi H. Self-reported and measured weight, height and body mass index (BMI) in Italy, the Netherlands and North America. The European Journal of Public Health. 2011;21(4):414–9. Kuczmarski MF, Kuczmarski RJ, Najjar M. Effects of age on validity of self-reported height, weight, and body mass index: findings from the Third National Health and Nutrition Examination Survey, 1988–1994. Journal of the American Dietetic Association. 2001;101(1):28–34. Averett S, Korenman S. The economic reality of the beauty myth. National Bureau of Economic Research Cambridge, Mass., USA; 1993. Morris S. Body mass index and occupational attainment. Journal of health economics. 2006;25(2):347–64. Kinge JM. Body mass index and employment status: a new look. Economics & Human Biology. 2016;22:117–25. Campbell DD, Green M, Davies N, Demou E, Ward J, Howe LD, et al. Effects of increased body mass index on employment status: a Mendelian randomisation study. International Journal of Obesity. 2021:1–12. Edwards CH, Bjørngaard JH, Minet Kinge J. The relationship between body mass index and income: Using genetic variants from HUNT as instrumental variables. Health Economics. 2021;30(8):1933–49. Agerström J, Rooth D-O. The role of automatic obesity stereotypes in real hiring discrimination. Journal of Applied Psychology. 2011;96(4):790. Finkelstein LM, Frautschy Demuth RL, Sweeney DL. Bias against overweight job applicants: Further explorations of when and why. Human Resource Management: Published in Cooperation with the School of Business Administration, The University of Michigan and in alliance with the Society of Human Resources Management. 2007;46(2):203–22. Campos-Vazquez RM, Gonzalez E. Obesity and hiring discrimination. Economics & Human Biology. 2020;37:100850. Rooth D-O. Obesity, attractiveness, and differential treatment in hiring a field experiment. Journal of human resources. 2009;44(3):710–35. Morris S. The impact of obesity on employment. Labour Economics. 2007;14(3):413–33. Balsa AI, McGuire TG. Prejudice, clinical uncertainty and stereotyping as sources of health disparities. Journal of health economics. 2003;22(1):89–116. Åsvold BO, Langhammer A, Rehn TA, Kjelvik G, Grøntvedt TV, Sørgjerd EP, et al. Cohort Profile Update: The HUNT Study, Norway. International Journal of Epidemiology. 2022. Norges Bank. Valutakurser 2019 [Available from: https://www.norges-bank.no/tema/Statistikk/valutakurser/ . Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH, Day FR, et al. Genetic studies of body mass index yield new insights for obesity biology. Nature. 2015;518(7538):197–206. Brandkvist M, Bjørngaard JH, Ødegård RA, Åsvold BO, Sund ER, Vie GÅ. Quantifying the impact of genes on body mass index during the obesity epidemic: longitudinal findings from the HUNT Study. bmj. 2019;366:l4067. Paternoster L, Tilling K, Davey Smith G. Genetic epidemiology and Mendelian randomization for informing disease therapeutics: Conceptual and methodological challenges. PLoS genetics. 2017;13(10):e1006944. Brandkvist M, Bjørngaard JH, Ødegård RA, Åsvold BO, Smith GD, Brumpton B, et al. Separating the genetics of childhood and adult obesity: a validation study of genetic scores for body mass index in adolescence and adulthood in the HUNT Study. Human molecular genetics. 2020;29(24):3966–73. Kent S, Fusco F, Gray A, Jebb SA, Cairns BJ, Mihaylova B. Body mass index and healthcare costs: a systematic literature review of individual participant data studies. Obesity Reviews. 2017;18(8):869–79. Staiger DO, Stock JH. Instrumental variables regression with weak instruments. National Bureau of Economic Research Cambridge, Mass., USA; 1994. Stock JH, Yogo M. Testing for weak instruments in linear IV regression. National Bureau of Economic Research Cambridge, Mass., USA; 2002. Burgess S, Small DS, Thompson SG. A review of instrumental variable estimators for Mendelian randomization. Statistical methods in medical research. 2017;26(5):2333–55. Buscot M-J, Wu F, Juonala M, Lehtimäki T, Pitkänen N, Sabin MA, et al. Longitudinal association of a body mass index (BMI) genetic risk score with growth and BMI changes across the life course: The Cardiovascular Risk in Young Finns Study. International Journal of Obesity. 2020;44(8):1733–42. Edwards CH, Aas E, Kinge JM. Body mass index and lifetime healthcare utilization. BMC health services research. 2019;19(1):1–10. Statistics Norway. Likestillingsutfordringer i deltidsarbeid og utdanningsnivå 2022 [10.06.2022]. Available from: https://www.ssb.no/befolkning/likestilling/statistikk/indikatorer-for-kjonnslikestilling-i-kommunene/artikler/likestillingsutfordringer-i-deltidsarbeid-og-utdanningsniva . Ulset E, Undheim R, Malterud K. Er fedmeepidemien kommet til Norge? Tidsskrift for Den norske legeforening. 2007. Table Table 1 is available in the Supplementary Files section Additional Declarations There is NO conflict of interest to disclose Supplementary Files SupplementaryInformation.pdf Table1.xlsx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-2275770","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":152265170,"identity":"f18a4878-0ebe-46df-9b01-993d5f103e5b","order_by":0,"name":"Christina Edwards","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABFUlEQVRIiWNgGAWjYDACCQaGA2CSgQeI2UBCCYwPCGtJQNXCbMCQgF8LA0QBQgubBD4t/LN7Hx78+cMij7+B9wDDh7LD8ubt6c+qC3/cSWyQbj6A1ZI7xw0O8yRIFEsc4EtgnHHusOGcM2/Mbs9IeJbYIHMMq10GEmkMh4F+SWw4wGPAzNt2m3GGRA7bbZ6Ew4kNEjkGuLQc/AHUMh+k5W/bbfsZEunPiglpOQB0WOIGkBbGttuJMyQSzJjxaZG4AXQYT5pE4sbDPAYHe879T57B88ZYmiftsHGbRBpWv/DPSGP++MOmLnHe8R7DBz/K0mxnsKc//Mxjc1i2XyIZa4ghADM4GSABNvzqR8EoGAWjYBTgAQDqz2HcjhxvFwAAAABJRU5ErkJggg==","orcid":"","institution":"Norwegian University of Science and Technology","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Christina","middleName":"","lastName":"Edwards","suffix":""},{"id":152265171,"identity":"767fe9f4-9203-41a0-bc68-b482e9ddc812","order_by":1,"name":"Johan H. Bjørngaard","email":"","orcid":"","institution":"Institutt for laboratoriemedisin, barne- og kvinnesykdommer","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Johan","middleName":"H.","lastName":"Bjørngaard","suffix":""},{"id":152265172,"identity":"e84a3609-21dd-4247-b678-da9dabd23a6e","order_by":2,"name":"Jonas Kinge","email":"","orcid":"","institution":"","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Jonas","middleName":"","lastName":"Kinge","suffix":""}],"badges":[],"createdAt":"2022-11-15 10:50:46","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-2275770/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-2275770/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":29253916,"identity":"c96af76c-cbb3-4bcd-99b4-a2304100d390","added_by":"auto","created_at":"2022-11-18 18:17:19","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":130953,"visible":true,"origin":"","legend":"\u003cp\u003eEffect (and 95 % CI) of BMI on probability of earning an income (upper panels) and income (2016 €) (lower panels), for males (left panels) and females (right panels) at ages 20 to 70, estimated using multivariable regression analyses (orange), and instrumental variable analysis with a weighted GRS as the instrument (green).\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-2275770/v1/f94c92596f79faa3f2362e4b.png"},{"id":29253917,"identity":"23b54fcb-a474-4af5-aab1-c165d9f67d9c","added_by":"auto","created_at":"2022-11-18 18:17:19","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":102438,"visible":true,"origin":"","legend":"\u003cp\u003eEffect (and 95% CI) (2016 €) of BMI on average income (2016 €) at ages 20-29, 30-39, 40-49, 50-59, and 60-69, estimated using multivariable adjusted analysis (top panel), and instrumental variable analysis (bottom panel) for male (left panels) and females (right panels).\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-2275770/v1/e7f8adddea6baffff246b8f2.png"},{"id":30004502,"identity":"5706dbcd-3ca6-4dc1-8600-8e327907242d","added_by":"auto","created_at":"2022-12-07 10:32:14","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":664437,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-2275770/v1/3d1fb2d3-0e75-452a-bb7f-3da3e23807a4.pdf"},{"id":29253919,"identity":"49402f45-7203-4d04-9434-3d7b3f420cd0","added_by":"auto","created_at":"2022-11-18 18:17:19","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1828086,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"SupplementaryInformation.pdf","url":"https://assets-eu.researchsquare.com/files/rs-2275770/v1/e71ed81c08871b98729c2c2a.pdf"},{"id":29253918,"identity":"b9d0586a-2c11-4ccb-ae40-3a1d72219477","added_by":"auto","created_at":"2022-11-18 18:17:19","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":13317,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"Table1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-2275770/v1/eef64ea329a0f83687609a25.xlsx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e conflict of interest to disclose","formattedTitle":"Obesity and sex- and age-specific income – evidence from the HUNT study","fulltext":[{"header":"Introduction","content":"\u003cp\u003ePast studies have found that increased body mass index (BMI) leads to lower income for females, and that male income seems less affected, or unaffected, by BMI (\u003cspan additionalcitationids=\"CR2 CR3 CR4 CR5\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). However, whether the effect of BMI on income differs by age is not known.\u003c/p\u003e \u003cp\u003eStudying the effects of BMI on income throughout a work life can give new insights into how BMI influences income. Adolescent obesity has been found to lead to reduced earnings in adulthood (age 24\u0026ndash;36), especially for females (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). If wages or the probability of income is affected early in adulthood, the effects may influence earnings throughout work life. On the other hand, one may expect that increased BMI is less likely to influence income in early adulthood, because the risk of BMI-related morbidities is smaller in younger adults (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e) and time lived with obesity is shorter. As age increases, so does the likelihood of ill health due to elevated BMI, which may influence income through increased absenteeism, presenteeism, job opportunities, and/or labour market participation.\u003c/p\u003e \u003cp\u003eA central methodological challenge has been to adequately address endogeneity biasing the relationship between BMI and income. BMI may lead to increased morbidity and mortality, which may reduce work ability and thereby wages. BMI could also influence income via, for example, employment, job promotion, choice of occupation, or job placements. On the other hand, income may influence accessibility to food and exercise options, which could influence weight (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). Second, the association between BMI and income may be confounded by a multitude of measurable and unmeasurable variables. For example, personality traits may influence income and likelihood of weight gain. Third, measurement error in studies using self-reported height, weight, and income, may affect the estimated effect of BMI on income (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSeveral studies have tried to overcome these endogeneity problems by employing an IV approach using instruments such as: past BMI (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e), area-level BMI (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e), BMI of a biological relative (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e), or genetic variants associated with BMI (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). These studies have also typically found that there is an association between BMI and income for females, but not for males. The main explanations proposed for this finding are: that increased BMI is associated with poorer health and may lead to lower work ability and thus wages, that individuals with elevated BMI are discriminated against on the labour market (\u003cspan additionalcitationids=\"CR18 CR19\" citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e), and that there is uncertainty about whether employing workers with obesity will lead to increased costs and perhaps reduced productivity (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn this paper, we combined data on objectively measured height and weight, BMI-related genetic variants, and sociodemographic variables from a longitudinal health survey with registry data on income from tax records to study the effect of BMI on one-year age-specific income at ages 20\u0026ndash;70 using multivariable adjusted and IV analyses.\u003c/p\u003e"},{"header":"Materials And Methods","content":"\u003cp\u003eWe used data from two rounds of the Tr\u0026oslash;ndelag Health Studies (The HUNT studies). These population-based cohort studies were conducted in a geographical region in Central Norway between August 1995 and June 1997 (HUNT2) and between October 2006 and June 2008 (HUNT3). In these studies, the height and weight of participants was measured, and biological materials were collected for genotyping performed using Illumina HumanCoreExome arrays. Detailed information about The HUNT Studies is available from \u0026Aring;svold, Langhammer (\u003cspan class=\"CitationRef\"\u003e23\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eWe linked the data from the HUNT studies, using a de-identified key, to national data available from Statistics Norway: The Norwegian Tax Administration Database and The Norwegian Population Registry. The Norwegian Tax administration database contains information about income, defined as the sum of personal and entrepreneurial income, from 1967 to 2016. The Norwegian National Registry contains information about whether individuals were alive and living in Norway during the years between 1986 and 2016.\u003c/p\u003e\n\u003cp\u003eThe annual income reported for each year from 1967 to 2016 was adjusted to 2016 price levels using the Norwegian Consumer Price Index. We then calculated one-year age-specific annual historical income for ages 20 to 70, using participant\u0026rsquo;s birth year. For instance, to calculate income at age 30 for a person born in 1975, we used income measured in 2005. For income measured between 1986 and 2016, participants were excluded at the year of death or not living in Norway. For income measured before 1986, all individuals were included. Income was converted from Norwegian Kroner (NOK) to 2016 Euros (\u0026euro; 1\u0026thinsp;=\u0026thinsp;NOK 9.29) (\u003cspan class=\"CitationRef\"\u003e24\u003c/span\u003e).\u003c/p\u003e\n\u003cdiv class=\"Section2\" id=\"Sec3\"\u003e\n \u003ch2\u003eRegression analyses\u003c/h2\u003e\n \u003cp\u003eWe estimated the one-year age-specific effect of BMI on: i) annual probability of working, ii) annual income, and iii) log of annual income, iv) z-score of annual income v) annual income among workers, and vi) z-score of annual income among workers, for ages 20\u0026ndash;70. All analyses were stratified by sex, because the effect of BMI on labour market outcomes such as income differ by sex.\u003c/p\u003e\n \u003cp\u003eFirst, we conducted instrumental variable (IV) analyses using two-stage least squares (2SLS) regressions using a weighted allele score as an instrument for BMI. These analyses were adjusted for birth year and study period. The alleles used in the score were selected based on 97 genetic variants which have been found to be associated with BMI in GWAS studies (\u003cspan class=\"CitationRef\"\u003e25\u003c/span\u003e). All, except one (rs12016871) of these variants were available in our dataset, and for this variant we followed Brandkvist, Bj\u0026oslash;rngaard (\u003cspan class=\"CitationRef\"\u003e26\u003c/span\u003e) and used the variant rs4771122 as a proxy. The weighted allele score was calculated by summing the weights (beta-coefficients reported by Locke and co-workers\u0026rsquo;) of the BMI-increasing alleles for each participant.\u003c/p\u003e\n \u003cp\u003eFor each one-year age group we first estimated the effect of the instrument (Z) on BMI (Eq.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e)\u003c/p\u003e\n \u003cdiv class=\"Equation\" id=\"Equ1\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e$${BMI}_{i}={\\pi }_{0}+ {\\pi }_{1}{Z}_{i} + {\\pi }_{2}{Sex}_{i}+ {\\pi }_{3}{birthyear}_{i}+ {\\pi }_{4}{study}_{i}+ {v}_{i}$$\u003c/div\u003e\n \u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\n \u003c/div\u003e\n \u003cp\u003eThen, using the predicted values from the first-stage (Eq. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e) we estimated the effect of BMI on each of the income-related outcomes (I\u003csub\u003e(1-6)\u003c/sub\u003e: I\u003csub\u003e1\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;probability of working, I\u003csub\u003e2\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;annual income, I\u003csub\u003e3\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;log of income, I\u003csub\u003e4\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;Income z-score, I\u003csub\u003e5\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;income among workers, I\u003csub\u003e6\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;z-score of income among workers) for each age-group (Eq.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e)\u003c/p\u003e\n \u003cdiv class=\"Equation\" id=\"Equ2\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e$${\\text{I}}_{({1-6)}_{i}}={\\beta }_{0}+ {\\beta }_{1}{\\widehat{BMI}}_{i}+ {\\beta }_{2}{Sex}_{i}+ {\\beta }_{3}{birthyear}_{i}+ {\\beta }_{4}{study}_{i}+ {u}_{i}$$\u003c/div\u003e\n \u003cdiv class=\"EquationNumber\"\u003e2\u003c/div\u003e\n \u003c/div\u003e\n \u003cp\u003eThe likelihood of working was estimated as the likelihood of having an income above zero. Log(income) was computed at each age as the log of income, where 0 income was set as 0.01. The income z-score was calculated as z = (x-\u0026micro;)/\u0026sigma;, where \u0026micro; and \u0026sigma; were calculated within each one-year age group for males and females. As expected, some participants had zero income despite being alive and living in the country. We therefore conducted a set of analyses including only participants that were working (i.e., with a positive income). Such analyses may be informative but should be interpreted with caution because we are conditioning on a possible consequence of the exposure (\u003cspan class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e28\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eNext, each of the outcomes (I\u003csub\u003e1-6\u003c/sub\u003e) were analysed using multivariable analyses adjusted for sex, birth year, study period when BMI measure was taken (HUNT2 or HUNT3), marital status (married, unmarried, divorced, widowed), smoking status (former smoker, sometimes smoker, daily smoker, and never smoker), educational level (primary school, secondary school, higher education), and urbanity (rural or urban). These covariates were selected because they have commonly been used in previous literature (\u003cspan class=\"CitationRef\"\u003e29\u003c/span\u003e).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"Section2\" id=\"Sec4\"\u003e\n \u003ch2\u003eInstrumental variable assumptions and non-linear associations\u003c/h2\u003e\n \u003cp\u003eThe instrumental variable assumptions and non-linearity in the relationship between BMI and income has been explored in a previous paper assessing the sex-specific effects of BMI on income using a similar sample, but without age-specific estimates (\u003cspan class=\"CitationRef\"\u003e16\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eFor the IV analyses to provide consistent results, the IV-assumptions must be satisfied. There are three main assumptions.\u003c/p\u003e\u003cspan\u003e\n \u003cp\u003e1. The weighted allele score must be associated with income (\u003cem\u003erelevance assumption\u003c/em\u003e). An F-statistic\u0026thinsp;\u0026gt;\u0026thinsp;10 is typically considered satisfactory (\u003cspan class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e31\u003c/span\u003e), but does not guarantee that any specific estimate will be unbiased or sufficiently well powered for the hypothesis of interest (\u003cspan class=\"CitationRef\"\u003e32\u003c/span\u003e).\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003e2. The weighted allele score should not be associated with any confounders (measured or unmeasured) (\u003cem\u003eindependence assumption\u003c/em\u003e).\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003e3. The weighted allele score should only influence income via BMI (\u003cem\u003eexclusion restriction\u003c/em\u003e).\u003c/p\u003e\n \u003c/span\u003e\n \u003cp\u003eIn the study assessing the non-age specific effect of BMI on income, the validity of the IV-assumptions was investigated using the weighted median, weighted modal, and MR-Egger approaches, as well as within-family analyses (\u003cspan class=\"CitationRef\"\u003e16\u003c/span\u003e). The findings suggested that there was little evidence of the issues typically associated with breaches in the IV-assumptions in studies using genetic variants as IVs (i.e., there was little evidence of directional pleiotropy, population stratification, assortative mating and dynastic effects). Nevertheless, we cannot be certain that assumptions 2 and 3 are fulfilled. We report F-statistics for each analysis at each age, to test the relevance assumption. Previous studies have found a stable effect from this allele score in different ages (\u003cspan class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e33\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eThere was no evidence of nonlinearities, in males nor females, at normal to higher levels of BMI, though some evidence of nonlinearity at lower (underweight) levels of BMI was found for males. However, the statistical power may have been insufficient to detect an effect.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"Section2\" id=\"Sec5\"\u003e\n \u003ch2\u003eBirth year specific effects\u003c/h2\u003e\n \u003cp\u003eAs a sensitivity analysis, we estimated birth year-specific effects. The aim of these analyses was to try to uncover generational effects. Here we calculated the mean income for ten-year age groups: 20\u0026ndash;29, 30\u0026ndash;39, 40\u0026ndash;49, 50\u0026ndash;59, and 60\u0026ndash;69. Then we estimated the sex-specific effects of BMI on mean ten-year income, for participants born in or before 1940, and individuals born after 1940. We applied the same methods as described above. We did not stratify into finer birth year groups to avoid losing statistical power.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Results","content":"\u003cdiv class=\"Section2\" id=\"Sec7\"\u003e\n \u003ch2\u003eDescriptive information\u003c/h2\u003e\n \u003cp\u003eThe average age of participants at the time that BMI was measured was 53 (SD 17.1, range: 19\u0026ndash;99) for males, and 53 (SD 17.7, range:19\u0026ndash;96) for females. The average BMI was 27.1 (SD 3.8, range:15\u0026ndash;53) for males, and 26.9 (SD 4.9, range:12\u0026ndash;56) for females. The majority of males were. classified as overweight, born between 1950 and 1959, married or cohabitating, had a secondary school education, lived in urban areas, and were never or former smokers (Table\u0026nbsp;1). The majority of females were: classified as normal weight or overweight, were born between 1950 and 1959, were married or cohabitating, had a secondary school education, lived in urban areas, and were never smokers (Table\u0026nbsp;1). Plots showing how these variables changed by birth year can be found in the Supplementary Information Appendix (SI Appendix) (Fig. S1, Fig. S2, and Fig. S3).\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eRegression Results\u003c/h3\u003e\n\u003cdiv class=\"Section2\" id=\"Sec9\"\u003e\n \u003ch2\u003eResults for males\u003c/h2\u003e\n \u003cp\u003eAt the start of working life, males seemed to benefit slightly from elevated BMI (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). From age 22 until 49, the effect of BMI on income was small, and the association was both positive and negative. From age 49 onwards, however elevated BMI was negatively associated with income.\u003c/p\u003e\n \u003cp\u003eThe results from the IV-models indicated that the probability of earning an income decreased with increasing BMI from age 22 onwards (except at ages: 24, 35 ,40\u0026ndash;43 ,69, and 70). The multivariable adjusted results coincided with the results from the IV analyses, but at most ages the coefficients estimated from the IV analysis were larger and less precise (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e, SI Appendix, Table S1).\u003c/p\u003e\n \u003cp\u003eThe effect of BMI on income followed a similar trend to the association between BMI and the probability of earning an income (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). The results from the IV-analyses indicated that from age 20\u0026ndash;25 the effect of BMI on income was positive (range: 56.4 to 232.9, mean\u0026thinsp;=\u0026thinsp;128.7), from age 26\u0026ndash;30 the effect estimates were negative (range: -63.6 to -189.1, mean = -113.9). Between age 30 and 48 the effect estimates varied between positive and negative (range: -177.5 to 134.9, mean =-6.2). From age 49 onwards, the effect estimate was negative except at ages 65 and 66 (range: -1295.9 to 93.5, mean = -533.0). The F-statistic was between 237 and 820, and declined with age (SI Appendix, Table S7). The findings from the multivariable adjusted analyses were similar, however, the effect estimates were smaller and more precise (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e, SI Appendix, Table S2).\u003c/p\u003e\n \u003cp\u003eWhen estimating the effect of BMI on log income (SI Appendix, Table S3, Fig. S4), the effects attenuated somewhat, but the age trend was similar to that of the effect of BMI on income. The mean z-score of income was \u0026minus;\u0026thinsp;0.008 (range: -0.044 to 0.014) in the IV analyses and was \u0026minus;\u0026thinsp;0.001 (range: -0.010 to 0.007) in the multivariable adjusted analyses (SI Appendix, Table S4). The income z-score and the conditional effect estimates (income given income and income z-score given income) followed a trend resembling that of the effect of BMI on income (SI Appendix, Table S5, Table S6, and Fig. S4).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"Section2\" id=\"Sec10\"\u003e\n \u003ch2\u003eResults for females\u003c/h2\u003e\n \u003cp\u003eFor females, the effects were more consistently negative across ages, with a tendency for increasing negative effects of BMI as age increased (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eThe results from the IV analyses indicated that there was a negative association between BMI and the likelihood of earning an income, except for between age 32 and 42 (in this age-interval the effect estimates were mainly positive) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e, SI Appendix, Table S8). The results from the multivariable adjusted analyses suggested that increased BMI was associated with a smaller likelihood of earning an income at all ages.\u003c/p\u003e\n \u003cp\u003eThe results from the IV-models indicated that the association between BMI and income was negative at all ages, except at ages 68\u0026ndash;70, where the effect was positive (mean: -236.2, range: -606.2 to 46.7) (SI Appendix, Table S9). The F-statistic was between 212 and 812 and decreased with age (SI Appendix, Table S14). The results from the multivariable adjusted analyses were similar, with smaller effects and higher precision. In the multivariable adjusted analysis BMI was associated with a lower income at all ages (mean = -144.1, range = -290.7 to -16.6)\u003c/p\u003e\n \u003cp\u003eWhen log income was used as the outcome measure, the effect estimates attenuated somewhat (SI Appendix, Table S10, Fig. S5), compared with when income was used. The mean z-score of income was \u0026minus;\u0026thinsp;0.007 (range: -0.013 to -0.002) in the multivariable adjusted analysis and \u0026minus;\u0026thinsp;0.012 (range: -0.027 to 0.007) in the 2SLS (SI Appendix, Table S11). The income z-score and the conditional effect estimates (income given income and income z-score given income) followed a similar trend to the effect of BMI on income (SI Appendix, Table S12, Table S13, and Fig. S5).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"Section2\" id=\"Sec11\"\u003e\n \u003ch2\u003eBirth year-specific estimates\u003c/h2\u003e\n \u003cp\u003eTo explore differences between age effects and generational effects, we estimated 10-year age-specific effects for males and females born before and after 1940. The estimates from the IV analysis suggest a slight tendency that the effect of BMI differs depending on the birth year category (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e, bottom panel). For males there was no clear trend: the confidence intervals overlapped, and the effect estimates at different ages were both larger and smaller in the older and younger groups. For females, there was a clear tendency for that the effect of BMI on income was more negative for females born after 1940, compared with the younger group. The difference was most prominent for mean income between age 30 and 39. The multivariable adjusted analysis estimates were similar for males and females born before and after 1940 (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e, top panel).\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eWe show that the relationship between BMI and income is negative throughout work life for females. For males, the effect of BMI on income was most prominent and negative at the end of working life (with the negative effects starting around age 49). No impact of BMI on income in males was found at younger ages. These findings were consistent across the income outcomes investigated in this study.\u003c/p\u003e \u003cp\u003eOur study demonstrates that increased BMI also influences male income, but that these effects tend to appear closer to the end of work life and are less precise. There may be several explanations for this. BMI may influence the health of males and females differently. A previous study using national Norwegian data, found that healthcare use increased substantially for males around age 50 and onwards (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e). A similar, but smaller increase was seen for females. This coincides with our study findings. In addition, there are other important differences between the sexes: males and females have different types of occupations, females are over twice as likely to work part time (16.7% of males and 37.3% of females work part time (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e), females have maternity leaves, and are more likely to take time off to care for children, and the educational levels of males and females have changed over time. Moreover, the interplay between male and female wages within households may be relevant.\u003c/p\u003e \u003cp\u003eThe positive association between BMI and income at the youngest ages could be due to selection into higher education. If individuals with a higher BMI are more likely to select into jobs that do not require a higher education, then we would expect those with a higher BMI to have a higher income early in work life.\u003c/p\u003e \u003cp\u003eWhen we stratified by birth year, we found that the effect of BMI on mean income from age 30\u0026ndash;39 (and to some extent age 40\u0026ndash;49) was positive for females born in 1940 or after, and negative for females born after 1940. This finding may have several explanations and should be explored further. First, labour market participation among females was lower before 1940 (SI Appendix, Fig. S6), and the most common occupations for females has also changed over time. In addition, ages 30\u0026ndash;39 and 40\u0026ndash;49 are ages that females are likely to be caring for children, and this may affect income. Rights and opportunities to work while caring for children has also changed considerably over the last decades. In Norway, the increase in population BMI seems to have begun in the 1960s for males and in the 1980s for females (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e). Those born before 1940 may therefore have been less exposed to the environmental causes of obesity. The obesogenic environment may have affected males and females differently. Another explanation is that BMI-related labour market discrimination increased as more females entered the labour market.\u003c/p\u003e \u003cp\u003eA strength of this study is that we have used registry-based income, and measured height and weight to compute BMI. However, using historical income to compute age-specific income, and the data informing BMI, genetic risk, and sociodemographic variables were measured either between 1995 and 1997 or between 2006 and 2008. This means that we are mixing cohorts and the time-of-measurement may bias the study findings. In addition, there may be a small degree of survival bias since individuals must have survived to participate in the study. The amount of bias from these factors may be different for males and females. We have tried to address this by conducting various sensitivity analyses, and by providing descriptive information at different ages and for different birth years (to the extent that his was possible) to improve transparency.\u003c/p\u003e \u003cp\u003eAnother limitation is that we have used the genetic variants presented by Locke, Kahali (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e). Other more recent studies have revealed additional variants, but these were not available in our data. We also did not explore the age-specific extent of pleiotropy (i.e., that the BMI-related genetic variants influence other phenotypic traits that influence income). However, this has previously been done for the effect of BMI on male and female income (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). Our results, do suggest, however that the effect of BMI-related genetic variants on income increases with age (SI Appendix, Fig. S7 and Fig. S8). Lastly, we studied personal wages, rather than household income, and we did not consider how partners\u0026rsquo; income could influence personal income.\u003c/p\u003e \u003cp\u003eOur findings show that BMI influences income for both males and females and demonstrate the importance of conducting age-specific analyses. An increased understanding of the underlying mechanisms driving the age-specific effects for males and females identified in our study could provide clues about how to address income disparities driven by BMI and about how to reduce the societal costs of increased population BMI. Future studies, with larger sample sizes, should consider looking at age-specific effects for different occupational groups, and should also explore the role of height and education in the association between BMI and income. In addition, the effect of BMI on household income should be investigated as well as the effect of individual BMI on spouses\u0026rsquo; income.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIncreased BMI leads to lower income for both sexes. As BMI increases, females have a lower probability of earning an income and reduced annual income at most ages. For males BMI-driven reductions in income were evident from around age 49.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eConflict of interest statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no conflicts of interest to declare.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Norwegian Research Council; Grant number: 295989.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Tr\u0026oslash;ndelag Health Study (HUNT) is a collaboration between HUNT Research Centre (Faculty of Medicine and Health Sciences, Norwegian University of Science and Technology NTNU), Tr\u0026oslash;ndelag County Council, Central Norway Regional Health Authority, and the Norwegian Institute of Public Health.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors: conceptualization, reviewing and editing of manuscript, CHE: Formal analyses, and writing of original draft, JHB \u0026amp; JMK: Funding acquisition.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data that support the findings of this study are available from third parties. Restrictions apply to the availability of these data, which were used under license for this study. After de-identification, the individual participant data that underlie the results reported in this article are available to researchers whose proposed use of the data has been reviewed and approved by the data owners and \u003cem\u003eRegional Committees\u003c/em\u003e for Medical and Health Research \u003cem\u003eEthics (\u003c/em\u003ehttps://.portalen.no\u003cem\u003e)\u003c/em\u003e.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBrunello G, d\u0026rsquo;Hombres B. Does body weight affect wages?: Evidence from Europe. Economics \u0026amp; Human Biology. 2007;5(1):1\u0026ndash;19.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCawley J. The impact of obesity on wages. Journal of Human resources. 2004;39(2):451\u0026ndash;74.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHan E, Norton EC, Stearns SC. Weight and wages: fat versus lean paychecks. Health economics. 2009;18(5):535\u0026ndash;48.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaum CL, Ford WF. The wage effects of obesity: a longitudinal study. Health economics. 2004;13(9):885\u0026ndash;99.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVillar JG, Quintana-Domeque C. Income and body mass index in Europe. Economics \u0026amp; Human Biology. 2009;7(1):73\u0026ndash;83.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmis JM, Hussey A, Okunade AA. Adolescent obesity, educational attainment and adult earnings. Applied Economics Letters. 2014;21(13):945\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThompson D, Edelsberg J, Colditz GA, Bird AP, Oster G. Lifetime health and economic consequences of obesity. Archives of internal medicine. 1999;159(18):2177\u0026ndash;83.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhan SS, Ning H, Wilkins JT, Allen N, Carnethon M, Berry JD, et al. Association of body mass index with lifetime risk of cardiovascular disease and compression of morbidity. JAMA cardiology. 2018;3(4):280\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eButland B, Jebb S, Kopelman P, McPherson K, Thomas S, Mardell J, et al. Tackling obesities: future choices-project report: Citeseer; 2007.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKrul AJ, Daanen HA, Choi H. Self-reported and measured weight, height and body mass index (BMI) in Italy, the Netherlands and North America. The European Journal of Public Health. 2011;21(4):414\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKuczmarski MF, Kuczmarski RJ, Najjar M. Effects of age on validity of self-reported height, weight, and body mass index: findings from the Third National Health and Nutrition Examination Survey, 1988\u0026ndash;1994. Journal of the American Dietetic Association. 2001;101(1):28\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAverett S, Korenman S. The economic reality of the beauty myth. National Bureau of Economic Research Cambridge, Mass., USA; 1993.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorris S. Body mass index and occupational attainment. Journal of health economics. 2006;25(2):347\u0026ndash;64.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKinge JM. Body mass index and employment status: a new look. Economics \u0026amp; Human Biology. 2016;22:117\u0026ndash;25.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCampbell DD, Green M, Davies N, Demou E, Ward J, Howe LD, et al. Effects of increased body mass index on employment status: a Mendelian randomisation study. International Journal of Obesity. 2021:1\u0026ndash;12.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEdwards CH, Bj\u0026oslash;rngaard JH, Minet Kinge J. The relationship between body mass index and income: Using genetic variants from HUNT as instrumental variables. Health Economics. 2021;30(8):1933\u0026ndash;49.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAgerstr\u0026ouml;m J, Rooth D-O. The role of automatic obesity stereotypes in real hiring discrimination. Journal of Applied Psychology. 2011;96(4):790.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFinkelstein LM, Frautschy Demuth RL, Sweeney DL. Bias against overweight job applicants: Further explorations of when and why. Human Resource Management: Published in Cooperation with the School of Business Administration, The University of Michigan and in alliance with the Society of Human Resources Management. 2007;46(2):203\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCampos-Vazquez RM, Gonzalez E. Obesity and hiring discrimination. Economics \u0026amp; Human Biology. 2020;37:100850.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRooth D-O. Obesity, attractiveness, and differential treatment in hiring a field experiment. Journal of human resources. 2009;44(3):710\u0026ndash;35.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorris S. The impact of obesity on employment. Labour Economics. 2007;14(3):413\u0026ndash;33.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBalsa AI, McGuire TG. Prejudice, clinical uncertainty and stereotyping as sources of health disparities. Journal of health economics. 2003;22(1):89\u0026ndash;116.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e\u0026Aring;svold BO, Langhammer A, Rehn TA, Kjelvik G, Gr\u0026oslash;ntvedt TV, S\u0026oslash;rgjerd EP, et al. Cohort Profile Update: The HUNT Study, Norway. International Journal of Epidemiology. 2022.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNorges Bank. Valutakurser 2019 [Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.norges-bank.no/tema/Statistikk/valutakurser/\u003c/span\u003e\u003cspan address=\"https://www.norges-bank.no/tema/Statistikk/valutakurser/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLocke AE, Kahali B, Berndt SI, Justice AE, Pers TH, Day FR, et al. Genetic studies of body mass index yield new insights for obesity biology. Nature. 2015;518(7538):197\u0026ndash;206.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrandkvist M, Bj\u0026oslash;rngaard JH, \u0026Oslash;deg\u0026aring;rd RA, \u0026Aring;svold BO, Sund ER, Vie G\u0026Aring;. Quantifying the impact of genes on body mass index during the obesity epidemic: longitudinal findings from the HUNT Study. bmj. 2019;366:l4067.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePaternoster L, Tilling K, Davey Smith G. Genetic epidemiology and Mendelian randomization for informing disease therapeutics: Conceptual and methodological challenges. PLoS genetics. 2017;13(10):e1006944.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrandkvist M, Bj\u0026oslash;rngaard JH, \u0026Oslash;deg\u0026aring;rd RA, \u0026Aring;svold BO, Smith GD, Brumpton B, et al. Separating the genetics of childhood and adult obesity: a validation study of genetic scores for body mass index in adolescence and adulthood in the HUNT Study. Human molecular genetics. 2020;29(24):3966\u0026ndash;73.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKent S, Fusco F, Gray A, Jebb SA, Cairns BJ, Mihaylova B. Body mass index and healthcare costs: a systematic literature review of individual participant data studies. Obesity Reviews. 2017;18(8):869\u0026ndash;79.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStaiger DO, Stock JH. Instrumental variables regression with weak instruments. National Bureau of Economic Research Cambridge, Mass., USA; 1994.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStock JH, Yogo M. Testing for weak instruments in linear IV regression. National Bureau of Economic Research Cambridge, Mass., USA; 2002.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBurgess S, Small DS, Thompson SG. A review of instrumental variable estimators for Mendelian randomization. Statistical methods in medical research. 2017;26(5):2333\u0026ndash;55.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBuscot M-J, Wu F, Juonala M, Lehtim\u0026auml;ki T, Pitk\u0026auml;nen N, Sabin MA, et al. Longitudinal association of a body mass index (BMI) genetic risk score with growth and BMI changes across the life course: The Cardiovascular Risk in Young Finns Study. International Journal of Obesity. 2020;44(8):1733\u0026ndash;42.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEdwards CH, Aas E, Kinge JM. Body mass index and lifetime healthcare utilization. BMC health services research. 2019;19(1):1\u0026ndash;10.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStatistics Norway. Likestillingsutfordringer i deltidsarbeid og utdanningsniv\u0026aring; 2022 [10.06.2022]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ssb.no/befolkning/likestilling/statistikk/indikatorer-for-kjonnslikestilling-i-kommunene/artikler/likestillingsutfordringer-i-deltidsarbeid-og-utdanningsniva\u003c/span\u003e\u003cspan address=\"https://www.ssb.no/befolkning/likestilling/statistikk/indikatorer-for-kjonnslikestilling-i-kommunene/artikler/likestillingsutfordringer-i-deltidsarbeid-og-utdanningsniva\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUlset E, Undheim R, Malterud K. Er fedmeepidemien kommet til Norge? Tidsskrift for Den norske legeforening. 2007.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Table","content":"\u003cp\u003eTable 1 is available in the Supplementary Files section\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-2275770/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-2275770/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eElevated body mass index (BMI) has been found to be associated with lower income, especially among women, and increasing evidence suggests that this association is causal. However, there is limited knowledge about the sex-specific effect of BMI on income at different ages. The relationship between BMI and income may change with age due to, for example, BMI-related morbidities or discrimination. The aim of this study was to investigate the sex-specific causal link between BMI and income at different ages.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe age-, and sex-specific effects were estimated using an instrumental variable approach with genetic variants as instruments (i.e., Mendelian randomisation) in an effort to deal with reverse causality and omitted variables that may bias the relationship between BMI and income. We also reduced measurement error by using registry-based income and clinically measured height and weight.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFindings\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eElevated BMI led to a reduced likelihood of working, and lower income. For females, increased BMI led to lower income throughout, and particularly at the end of, work life. For males, increased BMI led to lower income from age 49 onwards.\u003c/p\u003e","manuscriptTitle":"Obesity and sex- and age-specific income – evidence from the HUNT study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2022-11-18 18:17:14","doi":"10.21203/rs.3.rs-2275770/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d71b7745-7f20-4e85-bc7e-e492d2f79bd2","owner":[],"postedDate":"November 18th, 2022","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":16981324,"name":"Health sciences/Signs and symptoms"},{"id":16981325,"name":"Health sciences/Risk factors"}],"tags":[],"updatedAt":"2022-12-07T10:32:07+00:00","versionOfRecord":[],"versionCreatedAt":"2022-11-18 18:17:14","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-2275770","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-2275770","identity":"rs-2275770","version":["v1"]},"buildId":"FbvkV6FR0MCFSLy54lSbu","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}