Estimating the genetically predicted effects of lifestyle risk factors, educational attainment and Alzheimer’s disease liability on weight change during midlife
==================================================================================================================================================================

* Grace M. Power
* Jessica Tyrrell
* Apostolos Gkatzionis
* Si Fang
* Jon Heron
* George Davey Smith
* Tom G. Richardson

## Abstract

**Background** Weight change is a major indicator of adverse health outcomes. This study aims to examine factors contributing to weight change over a one-year interval in midlife.

**Methods** Observational and one-sample Mendelian randomisation (MR) analyses were conducted to estimate effects on weight change compared to one year previously (mean age: 56.5 years) using data from the UK Biobank study (n=453,169). Risk factors included alcohol consumption, smoking intensity, body mass index (BMI), educational attainment and Alzheimer’s disease liability.

**Results** Observational analyses indicated strong evidence of an association between greater educational attainment and family history of Alzheimer’s disease with weight loss. In contrast, smoking intensity and higher BMI were associated with weight gain. MR analyses were consistent with observational estimates for educational attainment and Alzheimer’s disease liability on weight loss and provided strong evidence of a genetically predicted effect between higher overall BMI and weight gain. Whilst there was little evidence of a genetically predicted effect between smoking intensity and weight change in those who had ever smoked, when stratified, smoking intensity was associated with weight loss in current smokers and weight gain in previous smokers. There was little evidence of an association between alcohol consumption and weight change in the one-year period. Inverse probability weighting was used to account for non-random selection on smoking and alcohol status and further stratification by smoking status, indicating that our results were largely robust to collider bias.

**Conclusions** Individuals who have been in education for longer, may have more opportunity to reduce their weight in midlife. The effect of Alzheimer’s disease liability on weight loss could be indicative of early signs of dementia. The relationship between smoking intensity and weight change is complex, reinforcing the importance of combining interventions aimed at controlling weight and smoking cessation among cigarette smokers.

**Key Messages**

*   Observational and Mendelian randomisation analyses used in this study provide consistent evidence that educational attainment and Alzheimer’s disease liability have an effect on weight loss over one year during midlife.

*   This suggests that individuals experiencing unintentional weight loss in midlife may be at high risk of developing Alzheimer’s disease in later life.

*   These results additionally demonstrate the value of Mendelian randomisation in minimising spurious associations obtained through observational research when assessing lifestyle risk factors prone to residual confounding, such as smoking intensity.

*   Our findings show that the relationship between smoking and weight change is complex, endorsing the notion that strategies to tackle weight control and smoking cessation should be combined.

Key words
*   weight change
*   midlife
*   Alzheimer’s disease
*   educational attainment
*   smoking intensity
*   alcohol consumption
*   lifestyle risk factors
*   observational epidemiology
*   genetic epidemiology

## Background

Weight change across the lifecourse can have major health implications (1–3). Weight gain in adulthood has been associated with increased risk of cardiometabolic disease, including type 2 diabetes and coronary heart disease (4), whereas weight loss in midlife may be an early indicator of dementia and mortality (5–9). Furthermore, lifestyle risk factors, including increased alcohol consumption and smoking intensity, have been shown to associate with weight change. Alcohol has been reported to increase weight as a consequence of excess calorie intake (10). Evidence concerning the relationship between smoking and weight change, however, is both conflicting and complex (11, 12). Such inconsistencies are likely due to other lifestyle factors, since physical inactivity, unhealthy diet and alcohol consumption are positively correlated with both adiposity and smoking behaviour (13). There is evidence indicating that smoking intensity reduces weight within current smokers, whilst having been a heavier smoker previously has been reported to lead to greater weight gain amongst previous smokers (14). Furthermore, higher body mass index (BMI) has been shown to influence lifetime smoking, smoking initiation and smoking heaviness (15). In addition, a 2013 systematic review comprising 289 observational studies assessing educational attainment and obesity, identified an inverse relationship in studies from higher-income countries (16). This review revealed an association between lower education and obesity, whilst other studies have revealed an association between higher educational attainment and reduced weight gain (17). Evident within this review is a sparsity of observational investigations that adjusted for potential confounders, further highlighting the inherent issues of residual confounding within observational research (18).

Mendelian randomisation (MR) is a technique that can be used to generate more reliable evidence about causal effects of exposures on disease outcomes compared with observational studies (19). This is achieved by harnessing genetic variants associated with modifiable exposures, an approach that can be more robust to confounding factors and reverse causation than naïve observational analyses (20, 21). MR has been successfully applied to estimate the genetically predicted effect of biological, behavioural and socioeconomic exposures on disease risk (22–25). MR investigations have also been extended to evaluate the genetically predicted effects of disease liability on outcomes (26), which includes diseases that typically have a late-onset in the lifecourse, such as dementia. Whilst latent liability to dementia can be thought of as probabilistically underlying the binary diagnosis measure, results are therefore interpreted on the liability scale (27, 28). Within the data employed for our investigation, participants are likely to be at different stages of disease development and some who would never develop disease may still show influences of liability on outcomes, emphasising the need for interpretation to be the effect of liability, not disease.

This study aims to estimate the observational and genetically predicted effects of alcohol intake, smoking, BMI, educational attainment and Alzheimer’s disease liability on weight change compared to one year previously during midlife, using data on participants from the UK Biobank (UKB) study (1). Initially, observational analyses will be conducted using ordinal logistic regression (OLR). MR analyses will then be applied to compare the effects of genetically predicted exposures with observational estimates. In doing so, we sought to develop a better understanding of the factors that contribute to weight change in midlife to improve effective strategies in the prevention of related adverse outcomes (29, 30).

## Methods

### Data resources

Data were collected from the UKB study; a population-based cohort aged between 40-69 years at recruitment, enrolled in the United Kingdom between 2006 and 2010 (1). In depth details on data collection within UKB during clinic visits and through electronic health records, as well as genotyping, quality control and imputation have been described in detail previously (31).

All observational exposure data were collected cross-sectionally (Table S1). Smoking intensity by current smokers was assessed by number of cigarettes per day. Alcohol intake by ever drinkers was assessed by units per day, which was calculated from the NHS estimates of pure alcohol consumption per day using data collected via the UKB (32). Weekly and monthly average intake was converted to daily intake by dividing by 7 and 30, respectively. Educational attainment was derived as a continuous score using the criteria described previously by the SSGAC consortium (33). Family history of Alzheimer’s disease was analysed as a binary variable indicating either or both parents reportedly having Alzheimer’s disease, as well as subsequent analyses of mothers and fathers separately. The outcome variable used in this investigation was obtained from participants being asked the question “Compared with one year ago, has your weight changed?” with the response options “Yes – lost weight”, “No – weight about the same” and “Yes – gained weight” which were coded 0, 1 and 2, respectively. We used to the variable containing data on previous and current smokers to stratify on smoking status in MR analyses.

### Genetic risk scores

Instruments for each exposure were constructed as genetic risk scores coded additively in UKB using the software PLINK (34). Genetic variants robustly associated with alcohol intake (35), BMI (36), educational attainment (37) and Alzheimer’s disease (38) were identified from large-scale genome-wide association studies which did not include the UKB study. Variants were selected based on P<5×10−8 and r2<0.001 using a reference panel of Europeans from the 1000 genomes project phase 3 to ensure they were independent (39). Influence on smoking is poorly captured by the observational variable indicating number of cigarettes smoked per day (40). Thus, for smoking intensity, we used a single variant as our instrument (rs16969968) located at the *CHRNA5* locus which has been shown to be robustly associated with smoking heaviness (41).

### Statistical analysis

OLR analyses were undertaken to calculate observational estimates of each of the five exposures in turn on weight change compared to one year previously. Analyses were adjusted for age, sex and the top ten principal components to account for residual population stratification in the UKB study.

Univariable MR analyses were conducted using individual-level data by applying OLR with the genetic instruments described above in turn on weight change compared to one year previously. MR analyses were adjusted for the same covariates as in the observational analyses. Using one-sample MR, as opposed to two-sample MR, allowed us to remove participants who had reportedly never smoked or who were never drinkers. This ensured analyses were conducted accurately on appropriate populations: ever smoked (n=150,812) and ever drunk alcohol (n=312,948), making the modelling more intuitive.

Previous evidence has found that the effect of ever smoking on BMI weakens as time since smoking cessation increases (41, 42). Therefore, to get an accurate estimate of smoking on weight related outcomes, analyses should be undertaken in current smokers as a separate group. Consequently, MR analyses of the relationship between the genetic instrument used for smoking intensity and weight change were completed after stratifying on current (n=33,098) and previous (n=117,714) smoking. Furthermore, conducting MR analyses in current smokers ensures comparisons can made between MR and observational estimates, since the observational smoking intensity variable is measured by number of cigarettes smoked per day and is therefore collected for current smokers only.

### Collider bias

We used Inverse Probability Weighting (IPW) to account for both (i) selection on smoking and alcohol use and (ii) stratification by smoking status, both of which may have induced collider bias into analyses. This was to ensure our estimates had accounted for unmeasured confounding and non-random selection (43). As illustrated in Figure 1, the exclusion of never-smokers (or never-drinkers) opens a pathway between the genetic instrument used and the outcome, weight change compared to one year previously, via both measured and unmeasured confounders (Figure 1A).

![Figure 1A:](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2021/07/30/2021.07.21.21260895/F1.medium.gif)

[Figure 1A:](http://medrxiv.org/content/early/2021/07/30/2021.07.21.21260895/F1)

Figure 1A: 
A directed acyclic graph indicating potential collider bias encountered in this study when selecting participants based on smoking and alcohol drinking status in the UK Biobank.

![Figure 1B:](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2021/07/30/2021.07.21.21260895/F2.medium.gif)

[Figure 1B:](http://medrxiv.org/content/early/2021/07/30/2021.07.21.21260895/F2)

Figure 1B: 
A directed acyclic graph indicating potential collider bias encountered in this study when stratifying on participants’ smoking status in the UK Biobank.

Likewise, stratification within the ever smokers’ group to investigate the relationship between previous or current smoking on weight change may also induce bias (Figure 1B).

IPW, an approach that uses weights to attempt to make selected participants a representative sample of the study population, has been shown to ameliorate bias (43, 44). Here, we explain how we have applied this method to investigate the relationship between smoking intensity and weight change. The first stage was additionally computed to assess the association between alcohol consumption and weight change in MR analysis.

IPW weights were derived from a logit model estimated on the whole population. The logit model included the collider (never vs ever smokers) as the binary outcome, and risk factors for smoking initiation as independent variables. These weights were then included in the analysis assessing the effect of the genetic instrument used for smoking intensity on weight change in ever smokers. A second logit model was estimated within the ever (previous vs current) smoking population. This included risk factors for current smoking. Weights from this second model were multiplied by the previous weights described above. These new weights were then included in the stratified analysis assessing the effect of the genetic instrument for smoking intensity on weight change within previous (n=106,666) and current (n=7,669) smokers. Each selected participant is now not only accounting for themselves but for those with similar characteristics who were not selected (45). Reduced sample sizes were a result of including additional variables in our models as data on all variables was necessary for analyses. The first stage was additionally computed to assess the association between alcohol consumption and weight change in MR analysis.

### Sensitivity analyses

Additional sensitivity analyses were conducted using MR to apply the pleiotropy robust weighted median and MR-Egger methods using the ‘TwoSampleMR’ R package (46–48). This was computed for the exposures measuring alcohol consumption, educational attainment, BMI and genetically predicted Alzheimer’s disease to examine the robustness of results to horizontal pleiotropy; the phenomenon where genetic variants influence multiple traits or disease outcomes via independent biological pathways. Pleiotropy robust methods were not undertaken for smoking intensity due to only using a single SNP instrument at the *CHRNA5* locus.

## Results

Observational analyses adjusted for age, sex and the top ten principal components to account for residual population stratification indicated strong evidence of association between smoking intensity in current smokers and weight gain (OR, 95% CI: 1.013, 1.010 to 1.016, P= 1×10−16), higher overall BMI and weight gain (OR, 95% CI: 1.113, 1.112 to 1.115, P<1×10−300), educational attainment and weight loss (OR, 95% CI: 0.969, 0.968 to 0.971, P<1×10−300) and Alzheimer’s disease in either or both parents and weight loss (OR, 95% CI: 0.970, 0.951 to 0.990, P=0.003) (Table 1).

Causal effect estimates from one-sample MR were similar to the observational estimates for most exposures (Figure 2). For instance, there was strong evidence of a genetically predicted effect between higher overall BMI and weight gain (OR, 95% CI: 1.065, 1.058 to 1.072, P=5×10−75), educational attainment and weight loss (OR, 95% CI: 0.956, 0.950 to 0.963, P=3×10−39) and Alzheimer’s disease liability and weight loss (OR, 95% CI: 0.985, 0.978 to 0.991, P=7×10−6). In contrast, MR analyses provided little evidence of an effect between the genetic instrument used for smoking intensity and weight change in those that had ever smoked (i.e., previous and current smokers) (OR, 95% CI: 1.006, 0.991 to 1.021, P=0.5). However, upon stratifying ever smokers into previous and current, there was evidence that smoking intensity has a genetically predicted effect on weight loss in current smokers (OR, 95% CI: 0.964, 0.933 to 0.995, P=0.02) and weight gain in previous smokers (OR, 95% CI: 1.017, 1.000 to 1.035, P=0.04), with strong evidence of effect modification (P=0.004). There was little evidence of a genetically predicted effect of alcohol intake on weight change (OR, 95% CI: 1.00, 0.993 to 1.007, P=0.9) (Table 2).

![Figure 2:](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2021/07/30/2021.07.21.21260895/F3.medium.gif)

[Figure 2:](http://medrxiv.org/content/early/2021/07/30/2021.07.21.21260895/F3)

Figure 2: 
A forest plot comparing results from one-sample mendelian randomization (yellow) and observational (blue) analyses with the outcome variable comprising three tiers coded 0 (“Yes – lost weight”, 1 (“No – weight about the same”) and 2 (“Yes – gained weight”). Dots filled-in indicate some to very strong statistical evidence of an association and the exposure variable Alzheimer’s disease presented on the continuous liability scale in one-sample MR and coded 0 (“Alzheimer’s disease in neither parent”) and 1 (“Alzheimer’s disease in either or both parents”) in observational analyses.

### Collider bias

IPW analyses were conducted for estimates that may be biased by non-random selection. Initially these were computed for estimates calculated within the ever smoker and alcohol consumer groups. There was evidence of an effect between the genetic instrument used for smoking intensity and weight change in those that had ever smoked (i.e., previous and current smokers) (OR, 95% CI: 1.009, 1.002 to 1.016, P=0.009) and alcohol consumption and weight change in those that had ever drunk alcohol (OR, 95% CI: 0.997, 0.989 to 1.004, P=0.371) (Table 3). Upon stratifying, there was strong evidence of an effect between smoking intensity and weight change in those that had previously smoked (OR, 95% CI: 1.019, 1.008 to 1.030, P=0.001) and little evidence between smoking intensity and weight change in those that were current smokers (OR, 95% CI: 0.988, 0.948 to 1.029, P=0.6) (Table 3). Including additional variables into IPW models reduced sample sizes and therefore impacted the precision of these estimates.

### Sensitivity analyses

Findings from the two-sample MR weighted median and MR-Egger method revealed similar trends to one-sample MR, suggesting that effects detected are unlikely to be prone to horizontal pleiotropy (Table 4). MR leave one out sensitivity analysis showed that when ‘rs12972156’, located at the *APOE* locus on chromosome 19, is left out, the effect estimates for genetically predicted Alzheimer’s disease liability on weight change compared to one year previously attenuated but still supported an effect on weight loss (OR, 95% CI: 0.993, 0.988 to 0.998, P=0.008).

## Discussion

In this study, results from observational analyses indicated strong evidence of an association between four exposures (smoking intensity, BMI, educational attainment and Alzheimer’s disease family history) and weight change compared to one year previously in midlife (mean age: 56.5 years). Conversely, there was little evidence of a relationship between alcohol intake and weight change within this lifecourse interval. Using one-sample MR, there was strong evidence of an effect between genetically predicted BMI and weight gain, whereas genetically predicted educational attainment and Alzheimer’s disease liability provided evidence of an effect on weight loss. MR analyses found weak evidence of a genetically predicted effect of smoking intensity and alcohol consumption on weight change compared to one year previously in midlife, however, when those who had ever smoked were stratified by previous and current smoking, smoking intensity had a genetically predicted effect on weight gain and weight loss, respectively. Results from IPW analyses, conducted to account for potential collider bias, reveal consistent directions of effect with stronger evidence and weaker statistical evidence, respectively. This is likely a reflection of sample size reduction in the latter.

Educational attainment in our study was associated with weight loss in midlife based on evidence from both observational and MR analyses. An inverse association between educational attainment and obesity has been consistently observed in studies conducted in high-income countries (16). For instance, a prospective cohort study comprising children (n=12,150) born in Aberdeen between 1950 and 1956 concluded educational attainment was inversely associated with adult BMI (mean age 48 years) after controlling for childhood intelligence and other measured covariates (49). Findings from this study were explained by fixed family factors, such as parental socioeconomic background, intelligence and education, and school and neighbourhood characteristics. Results for educational attainment in our study may also be partially interpreted as a proxy for socioeconomic position. In addition to overall BMI, higher educational attainment, has been associated with a lower increase in weight gain during adulthood in a multicentre cohort study conducted between 1992 to 2000 across Europe (n=361,467) (17). Education, as well as other socioeconomic factors, may influence weight-related behaviours such as diet and physical activity, thus affecting energy balance. If health behaviours of this kind persist, this imbalance has the potential to accumulate over time, leading to accelerated increase in weight, and therefore, continued and greater weight gain in midlife.

Alzheimer’s disease liability was a strong predictor of weight loss in both observational and MR analyses. Importantly, using the *APOE* variant as a genetic instrument for Alzheimer’s disease liability may be less prone to misclassification than reported family history, which may explain the narrower confidence intervals when comparing observational and MR estimates for this exposure. This result was supported using a two-sample MR leave one out analysis undertaken. Here we found removal of the ‘rs12972156’ genetic variant, located at the *APOE* locus, attenuated the association between Alzheimer’s disease liability and weight loss. Nevertheless, findings consistently show a genetically predicted effect between Alzheimer’s disease liability and weight loss in midlife. These results are in line with previous observational research revealing that weight loss precedes the diagnosis of dementia (5) and MR research indicating an association between a higher polygenic risk score for Alzheimer’s disease and lower BMI and body fat (50).

Findings related to alcohol consumption were not consistent with *a priori* expectations derived from previous studies assessing long-term weight change (10). Specifically, evidence from the literature suggesting that increase in alcohol use is associated with weight gain. This may be in part due to our outcome measuring weight change compared to just one year previously at midlife, assuming that on average, those that drink in this age group have been doing so for a considerable length of time over the lifecourse. Thus, weight change as a result of drinking is likely to have had a substantial effect already, resulting in evidence of an impact within one year challenging to ascertain. In addition, observational analyses indicated that increased smoking intensity resulted in weight gain. Whilst there was no evidence of an association between smoking intensity and weight change compared to one year previously in midlife within those who had ever smoked in one-sample MR, when this group was stratified by smokers who previously smoked and who currently smoke, results were in line with evidence pointing to an inverse association between current cigarette smoking and body weight, followed by weight gain after smoking cessation (11, 14, 51, 52). Again, it is important to reiterate that careful consideration is required if stratifying a population into subgroups when conducting MR analyses, as well as observational analyses, to evaluate when collider bias may have been introduced. We express the importance of adjusting for this using IPW (43). These findings also highlight the potential for MR to better mitigate against distortions resulting from residual confounding, often present in observational epidemiology (19). It has also been established that there is a bi-directional association between heavy smoking and BMI, outlining the added complexity of this relationship (13). Higher levels of BMI have been reported to increase risk of smoking uptake as well as intensity of smoking (13). This highlights the need to combine weight control and smoking cessation strategies within interventions aimed at tackling these important public health concerns.

The association between BMI and weight gain during one year in midlife in observational and one-sample MR analyses suggests that individuals with a higher BMI will typically continue to gain weight at this stage in the lifecourse. Reduced metabolism levels in those with higher BMI are likely to lead to low energy, which may discourage physical activity, further enhanced by midlife body changes as a result of aging (53).

### Strengths and limitations

Observational analyses have previously been conducted alongside MR to compare findings; however, this study is distinctive in that it assesses a time varying outcome; weight change compared to one year previously at midlife. This study also utilises additional sensitivity analyses through weighted median and MR-Egger methods in two-sample MR to ensure the robustness of univariable results to horizontal pleiotropy.

Important limitations exist in this investigation, however. Firstly, the outcome is a self-reported measure of perceived weight change over a one-year period. Individuals with certain characteristics, including higher overall BMI and educational attainment, may be more likely to misreport their weight, indicating the potential for recall bias and thus differential misclassification. For instance, those with higher educational history have been shown to be more likely to underreport their weight, which would result in overestimated effect sizes (54). Secondly, the outcome used is restricted since it does not demonstrate weight change over a substantial period of time (i.e., more than one year). As discussed previously, this was a potential issue when assessing alcohol consumption, however, additionally has an added benefit in being able to examine recent and rapid weight change during midlife. Additionally, we do not have data on whether an individual changed their smoking or drinking throughout the year in which we examine their weight change. Another key limitation of this study is selection bias and thus limited generalisability. Participants in the UKB are older, more likely to be female and less likely to live in socially deprived areas, compared to participants in nationally representative data sources (55). In addition, previous MR studies have revealed that longer educational duration and older menarche have increased participation, whilst a genetic liability of neuroticism, Alzheimer’s disease and schizophrenia have reduced participation, providing further evidence for participation bias (56). Furthermore, (i) selection on smoking and alcohol use, and (ii) stratification by smoking status, may have induced collider bias (43, 57). Whilst IPW was conducted to account for this, including additional variables to generate weights resulted in smaller sample sizes and thus reduced power and precision in our IPW estimates.

## Conclusion

This investigation provides novel evidence of the effects of lifestyle risk factors, educational attainment and Alzheimer’s disease liability on weight change compared to one year previously in midlife. Results suggest that individuals who attended higher education may have more opportunity to help reduce weight gain during this stage in the lifecourse. At the same time, findings may have implications for early testing for Alzheimer’s disease when weight loss is observed at midlife. In addition, the relationship between smoking and weight change is complex with findings endorsing the idea that strategies to tackle weight control and smoking cessation should be combined.

## Supporting information

Tables [[supplements/260895_file02.xlsx]](pending:yes)

## Data Availability

The data that support the findings of this study are all available from within the supplementary materials or publicly accessible from the resources cited.

## Acknowledgments

We would like to thank the UK Biobank study and all participants who contributed to it, as well as the authors of all the genome-wide association studies who made their summary statistics available for the benefit of this work.

## Abbreviations

BMI
:   Body mass index
IPW
:   Inverse Probability Weighting
MR
:   Mendelian randomisation
OLR
:   Ordinal logistic regression
UKB
:   UK Biobank

*   Received July 21, 2021.
*   Revision received July 21, 2021.
*   Accepted July 30, 2021.


*   © 2021, Posted by Cold Spring Harbor Laboratory

This pre-print is available under a Creative Commons License (Attribution-NonCommercial-NoDerivs 4.0 International), CC BY-NC-ND 4.0, as described at [http://creativecommons.org/licenses/by-nc-nd/4.0/](http://creativecommons.org/licenses/by-nc-nd/4.0/)

## References

1.  1.Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, et al. UK Biobank: An Open Access Resource for Identifying the Causes of a Wide Range of Complex Diseases of Middle and Old Age. PLOS Medicine. 2015;12(3):e1001779.
    
    

2.  2.Yarnell JWG, Patterson CC, Thomas HF, Sweetnam PM. Comparison of weight in middle age, weight at 18 years, and weight change between, in predicting subsequent 14 year mortality and coronary events: Caerphilly Prospective Study. Journal of Epidemiology and Community Health. 2000;54(5):344.
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoiamVjaCI7czo1OiJyZXNpZCI7czo4OiI1NC81LzM0NCI7czo0OiJhdG9tIjtzOjUwOiIvbWVkcnhpdi9lYXJseS8yMDIxLzA3LzMwLzIwMjEuMDcuMjEuMjEyNjA4OTUuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 

3.  3.Corrada MM, Kawas CH, Mozaffar F, Paganini-Hill A. Association of Body Mass Index and Weight Change with All-Cause Mortality in the Elderly. American Journal of Epidemiology. 2006;163(10):938–49.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/aje/kwj114&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=16641311&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000237319100009&link_type=ISI) 

4.  4.Wilson PW, Kannel WB. Obesity, diabetes, and risk of cardiovascular disease in the elderly. Am J Geriatr Cardiol. 2002;11(2):119–23,25.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1111/j.1076-7460.2002.00998.x&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=11872970&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

5.  5.Knopman DS, Edland SD, Cha RH, Petersen RC, Rocca WA. Incident dementia in women is preceded by weight loss by at least a decade. Neurology. 2007;69(8):739–46.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1212/01.wnl.0000267661.65586.33&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=17709705&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

6.  6.Johnson DK, Wilkins CH, Morris JC. Accelerated Weight Loss May Precede Diagnosis in Alzheimer Disease. Archives of Neurology. 2006;63(9):1312–7.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1001/archneur.63.9.1312&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=16966511&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000240435500016&link_type=ISI) 

7.  7.Chen C, Ye Y, Zhang Y, Pan XF, Pan A. Weight change across adulthood in relation to all cause and cause specific mortality: prospective cohort study. BMJ. 2019;367:5584.
    
    

8.  8.Hamer M, O’Donovan G. Sarcopenic obesity, weight loss, and mortality: the English Longitudinal Study of Ageing. Am J Clin Nutr. 2017;106(1):125–9.
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NDoiYWpjbiI7czo1OiJyZXNpZCI7czo5OiIxMDYvMS8xMjUiO3M6NDoiYXRvbSI7czo1MDoiL21lZHJ4aXYvZWFybHkvMjAyMS8wNy8zMC8yMDIxLjA3LjIxLjIxMjYwODk1LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

9.  9.Wong CJ. Involuntary weight loss. Med Clin North Am. 2014;98(3):625–43.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.mcna.2014.01.012&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=24758965&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

10. 10.Mozaffarian D, Hao T, Rimm EB, Willett WC, Hu FB. Changes in diet and lifestyle and long-term weight gain in women and men. N Engl J Med. 2011;364(25):2392–404.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1056/NEJMoa1014296&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=21696306&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000291936100005&link_type=ISI) 

11. 11.Sneve M, Jorde R. Cross-sectional study on the relationship between body mass index and smoking, and longitudinal changes in body mass index in relation to change in smoking status: the Tromso Study. Scand J Public Health. 2008;36(4):397–407.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1177/1403494807088453&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=18539694&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000257536300009&link_type=ISI) 

12. 12.Dare S, Mackay DF, Pell JP. Relationship between smoking and obesity: a cross-sectional study of 499,504 middle-aged adults in the UK general population. PLoS One. 2015;10(4):e0123579.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1371/journal.pone.0123579&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=25886648&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

13. 13.Carreras-Torres R, Johansson M, Haycock PC, Relton CL, Davey Smith G, Brennan P, et al. Role of obesity in smoking behaviour: Mendelian randomisation study in UK Biobank. BMJ. 2018;361:k1767–k.
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiYm1qIjtzOjU6InJlc2lkIjtzOjE3OiIzNjEvbWF5MTVfOC9rMTc2NyI7czo0OiJhdG9tIjtzOjUwOiIvbWVkcnhpdi9lYXJseS8yMDIxLzA3LzMwLzIwMjEuMDcuMjEuMjEyNjA4OTUuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 

14. 14.Piirtola M, Jelenkovic A, Latvala A, Sund R, Honda C, Inui F, et al. Association of current and former smoking with body mass index: A study of smoking discordant twin pairs from 21 twin cohorts. PLOS ONE. 2018;13(7):e0200140.
    
    

15. 15.Taylor AE, Richmond RC, Palviainen T, Loukola A, Wootton RE, Kaprio J, et al. The effect of body mass index on smoking behaviour and nicotine metabolism: a Mendelian randomization study. Human molecular genetics. 2019;28(8):1322–30.
    
    

16. 16.Cohen AK, Rai M, Rehkopf DH, Abrams B. Educational attainment and obesity: a systematic review. Obes Rev. 2013;14(12):989–1005.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1111/obr.12062&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=23889851&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

17. 17.Rohrmann S, Steinbrecher A, Linseisen J, Hermann S, May A, Luan J, et al. The association of education with long-term weight change in the EPIC-PANACEA cohort. Eur J Clin Nutr. 2012;66(8):957–63.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/ejcn.2012.55&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=22669330&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000307236900013&link_type=ISI) 

18. 18.Smith GD, Ebrahim S. Data dredging, bias, or confounding. BMJ. 2002;325(7378):1437–8.
    
    [FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiRlVMTCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiYm1qIjtzOjU6InJlc2lkIjtzOjEzOiIzMjUvNzM3OC8xNDM3IjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjEvMDcvMzAvMjAyMS4wNy4yMS4yMTI2MDg5NS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

19. 19.Davey Smith G, Ebrahim S. ‘Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease?*. International Journal of Epidemiology. 2003;32(1):1–22.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/ije/dyg070&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=12689998&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000182341300001&link_type=ISI) 

20. 20.Davey Smith G, Ebrahim S. What can mendelian randomisation tell us about modifiable behavioural and environmental exposures? BMJ. 2005;330(7499):1076–9.
    
    [FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiRlVMTCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiYm1qIjtzOjU6InJlc2lkIjtzOjEzOiIzMzAvNzQ5OS8xMDc2IjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjEvMDcvMzAvMjAyMS4wNy4yMS4yMTI2MDg5NS5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

21. 21.Davey Smith G, Holmes MV, Davies NM, Ebrahim S. Mendel’s laws, Mendelian randomization and causal inference in observational data: substantive and nomenclatural issues. European Journal of Epidemiology. 2020;35(2):99–111.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/s10654-020-00622-7&link_type=DOI) 

22. 22.Swerdlow DI, Holmes MV, Kuchenbaecker KB, Engmann JE, Shah T, Sofat R, et al. The interleukin-6 receptor as a target for prevention of coronary heart disease: a mendelian randomisation analysis. Lancet. 2012;379(9822):1214–24.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S0140-6736(12)60110-X&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=22421340&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000302230400034&link_type=ISI) 

23. 23.Holmes MV, Dale CE, Zuccolo L, Silverwood RJ, Guo Y, Ye Z, et al. Association between alcohol and cardiovascular disease: Mendelian randomisation analysis based on individual participant data. BMJ. 2014;349:g4164.
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiYm1qIjtzOjU6InJlc2lkIjtzOjE3OiIzNDkvanVsMTBfNi9nNDE2NCI7czo0OiJhdG9tIjtzOjUwOiIvbWVkcnhpdi9lYXJseS8yMDIxLzA3LzMwLzIwMjEuMDcuMjEuMjEyNjA4OTUuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 

24. 24.Nguyen TT, Tchetgen Tchetgen EJ, Kawachi I, Gilman SE, Walter S, Liu SY, et al. Instrumental variable approaches to identifying the causal effect of educational attainment on dementia risk. Ann Epidemiol. 2016;26(1):71–6.e1-3.
    
    

25. 25.Tillmann T, Vaucher J, Okbay A, Pikhart H, Peasey A, Kubinova R, et al. Education and coronary heart disease: mendelian randomisation study. BMJ. 2017;358:j3542.
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiYm1qIjtzOjU6InJlc2lkIjtzOjE3OiIzNTgvYXVnMzBfNy9qMzU0MiI7czo0OiJhdG9tIjtzOjUwOiIvbWVkcnhpdi9lYXJseS8yMDIxLzA3LzMwLzIwMjEuMDcuMjEuMjEyNjA4OTUuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 

26. 26.Tudball MJ, Bowden J, Hughes RA, Ly A, Munafò MR, Tilling K, et al. Mendelian randomisation with coarsened exposures. Genet Epidemiol. 2021;45(3):338–50.
    
    

27. 27.Davey Smith G. Post–Modern Epidemiology: When Methods Meet Matter. American Journal of Epidemiology. 2019;188(8):1410–9.
    
    

28. 28.Edwards JH. Familial predisposition in man. Br Med Bull. 1969;25(1):58–64.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/oxfordjournals.bmb.a070672&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=5782760&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=A1969C363200010&link_type=ISI) 

29. 29.Strand BH, Wills AK, Langballe EM, Rosness TA, Engedal K, Bjertness E. Weight Change in Midlife and Risk of Mortality From Dementia up to 35 Years Later. The Journals of Gerontology: Series A. 2017;72(6):855–60.
    
    

30. 30.Janssen I, Bacon E. Effect of Current and Midlife Obesity Status on Mortality Risk in the Elderly. Obesity. 2008;16(11):2504–9.
    
    

31. 31.Bycroft C, Freeman C, Petkova D, Band G, Elliott LT, Sharp K, et al. The UK Biobank resource with deep phenotyping and genomic data. Nature. 2018;562(7726):203–9.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41586-018-0579-z&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=30305743&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

32. 32.NHS. Calculating Alcohol Units [Available from: [https://www.nhs.uk/live-well/alcohol-support/calculating-alcohol-units/](https://www.nhs.uk/live-well/alcohol-support/calculating-alcohol-units/).
    
    

33. 33.Lee JJ, Wedow R, Okbay A, Kong E, Maghzian O, Zacher M, et al. Gene discovery and polygenic prediction from a genome-wide association study of educational attainment in 1.1 million individuals. Nature Genetics. 2018;50(8):1112–21.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41588-018-0147-3&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=30038396&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

34. 34.Chang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience. 2015;4:7-.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1186/s13742-015-0047-8&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=25722852&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

35. 35.Liu M, Jiang Y, Wedow R, Li Y, Brazel DM, Chen F, et al. Association studies of up to 1.2 million individuals yield new insights into the genetic etiology of tobacco and alcohol use. Nature Genetics. 2019;51(2):237–44.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41588-018-0307-5&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

36. 36.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.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/nature14177&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=25673413&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

37. 37.Okbay A, Beauchamp JP, Fontana MA, Lee JJ, Pers TH, Rietveld CA, et al. Genome-wide association study identifies 74 loci associated with educational attainment. Nature. 2016;533(7604):539–42.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/nature17671&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=27225129&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

38. 38.Lambert JC, Ibrahim-Verbaas CA, Harold D, Naj AC, Sims R, Bellenguez C, et al. Meta-analysis of 74,046 individuals identifies 11 new susceptibility loci for Alzheimer’s disease. Nat Genet. 2013;45(12):1452–8.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/ng.2802&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=24162737&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

39. 39.Abecasis GR, Auton A, Brooks LD, DePristo MA, Durbin RM, Handsaker RE, et al. An integrated map of genetic variation from 1,092 human genomes. Nature. 2012;491(7422):56–65.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/nature11632&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=23128226&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000310434500030&link_type=ISI) 

40. 40.Munafò MR, Timofeeva MN, Morris RW, Prieto-Merino D, Sattar N, Brennan P, et al. Association Between Genetic Variants on Chromosome 15q25 Locus and Objective Measures of Tobacco Exposure. JNCI: Journal of the National Cancer Institute. 2012;104(10):740–8.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/jnci/djs191&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=22534784&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000304199600008&link_type=ISI) 

41. 41.Millard LAC, Munafò MR, Tilling K, Wootton RE, Davey Smith G. MR-pheWAS with stratification and interaction: Searching for the causal effects of smoking heaviness identified an effect on facial aging. PLoS Genet. 2019;15(10):e1008353.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

42. 42.Freathy RM, Kazeem GR, Morris RW, Johnson PC, Paternoster L, Ebrahim S, et al. Genetic variation at CHRNA5-CHRNA3-CHRNB4 interacts with smoking status to influence body mass index. Int J Epidemiol. 2011;40(6):1617–28.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/ije/dyr077&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=21593077&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000297868500024&link_type=ISI) 

43. 43.Hughes RA, Davies NM, Davey Smith G, Tilling K. Selection Bias When Estimating Average Treatment Effects Using One-sample Instrumental Variable Analysis. Epidemiology. 2019;30(3).
    
    

44. 44.Gkatzionis A, Burgess S. Contextualizing selection bias in Mendelian randomization: how bad is it likely to be? International Journal of Epidemiology. 2019;48(3):691–701.
    
    

45. 45.Nohr EA, Liew Z. How to investigate and adjust for selection bias in cohort studies. Acta Obstetricia et Gynecologica Scandinavica. 2018;97(4):407–16.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

46. 46.Hemani G, Zheng J, Elsworth B, Wade KH, Haberland V, Baird D, et al. The MR-Base platform supports systematic causal inference across the human phenome. Elife. 2018;7.
    
    

47. 47.Bowden J, Davey Smith G, Burgess S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol. 2015;44(2):512–25.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/ije/dyv080&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26050253&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

48. 48.Bowden J, Davey Smith G, Haycock PC, Burgess S. Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator. Genet Epidemiol. 2016;40(4):304–14.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1002/gepi.21965&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=27061298&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

49. 49.Lawlor DA, Clark H, Davey Smith G, Leon DA. Childhood intelligence, educational attainment and adult body mass index: findings from a prospective cohort and within sibling-pairs analysis. International Journal of Obesity. 2006;30(12):1758–65.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/sj.ijo.0803330&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=16552398&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000242337400010&link_type=ISI) 

50. 50.Korologou-Linden R, Howe LD, Millard LAC, Ben-Shlomo Y, Williams DM, Smith GD, et al. The causes and consequences of Alzheimer’s disease: phenome-wide evidence from Mendelian randomization. medRxiv. 2020:2019.12.18.19013847.
    
    

51. 51.Munafò MR, Tilling K, Ben-Shlomo Y. Smoking status and body mass index: a longitudinal study. Nicotine Tob Res. 2009;11(6):765–71.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/ntr/ntp062&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=19443785&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000266698600025&link_type=ISI) 

52. 52.Kaufman A, Augustson EM, Patrick H. Unraveling the Relationship between Smoking and Weight: The Role of Sedentary Behavior. J Obes. 2012;2012:735465.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1155/2012/735465&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=21961058&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

53. 53.Khandelwal S. Obesity in midlife: lifestyle and dietary strategies. Climacteric. 2020;23(2):140–7.
    
    

54. 54.Stommel M, Schoenborn CA. Accuracy and usefulness of BMI measures based on self-reported weight and height: findings from the NHANES & NHIS 2001-2006. BMC Public Health. 2009;9:421.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1186/1471-2458-9-421&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=19922675&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom) 

55. 55.Sekula P, Del Greco M F, Pattaro C, Köttgen A. Mendelian Randomization as an Approach to Assess Causality Using Observational Data. Journal of the American Society of Nephrology. 2016;27(11):3253–65.
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6ODoiam5lcGhyb2wiO3M6NToicmVzaWQiO3M6MTA6IjI3LzExLzMyNTMiO3M6NDoiYXRvbSI7czo1MDoiL21lZHJ4aXYvZWFybHkvMjAyMS8wNy8zMC8yMDIxLjA3LjIxLjIxMjYwODk1LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

56. 56.Tyrrell J, Zheng J, Beaumont R, Hinton K, Richardson TG, Wood AR, et al. Genetic predictors of participation in optional components of UK Biobank. Nat Commun. 2021;12(1):886.
    
    

57. 57.Munafò MR, Tilling K, Taylor AE, Evans DM, Davey Smith G. Collider scope: when selection bias can substantially influence observed associations. International Journal of Epidemiology. 2018;47(1):226–35.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/ije/dyx206&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29040562&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2021%2F07%2F30%2F2021.07.21.21260895.atom)