Alcohol causes an increased risk of head and neck but not breast cancer in individuals from the UK Biobank study: A Mendelian randomisation analysis

ABSTACTAlcohol intake and the risk of various types of cancers have been previously correlated. Correlation though does not always mean that a causal relationship between the two is present. Excessive alcohol consumption is also correlated with other lifestyle factors and behaviours, such as smoking and increased adiposity, that also affect the risk of cancer and make the identification and estimation of the causal effect of alcohol on cancer difficult. Here, using individual level data for 322,193 individuals from the UK Biobank, we report the observational and causal effects of alcohol consumption on types of cancer previously suggested as correlated to alcohol. Alcohol was observationally associated with cancers of the lower digestive system, head and neck and breast cancer. No associations were observed when we considered those keeping alcohol consumption below the recommended threshold of 14 units/week. When Mendelian randomisation was used to assess the causal effect of alcohol on cancer, we found that increasing alcohol consumption, especially above the recommended level, was causal to head and neck cancers but not breast cancer. Our results where replicated using a two sample MR method and data from the much larger COGS genome wide analysis of breast cancer. We conclude that alcohol is causally related to head and neck cancers, especially cancer of larynx, but the observed association with breast cancer are likely due to confounding. The suggested threshold of 14 units/week appears suitable to manage the risk of cancer due to alcohol.

Download Full-text

Effects of alcohol consumption in general, and wine in particular, on the risk of cancer development: a review

OENO One ◽

10.20870/oeno-one.2020.54.4.3569 ◽

2020 ◽

Vol 54 (4) ◽

pp. 813-832

Author(s):

Pierre-Louis Teissedre ◽

Zurine Rasines-Perea ◽

Jean-Claude Ruf ◽

Creina Stockley ◽

Arina Oana Antoce ◽

...

Keyword(s):

Breast Cancer ◽

Alcohol Consumption ◽

Alcoholic Beverages ◽

International Agency ◽

Upper Aerodigestive Tract ◽

Body Tissues ◽

Increased Risk ◽

Different Types ◽

Lifestyle Pattern ◽

Risk Of Cancer

Since 1988, alcohol has been classified as a Group 1 carcinogen, the highest level of risk, by the International Agency for Research on Cancer (IARC). In fact, alcohol consumption is the third leading risk factor for disease and mortality in Europe. It accounts for 4.65 % of the global burden of both injury and disease, making it one of the most preventable causes of injury and death. Tissues in closest contact with alcohol when it is ingested, such as those of the oral cavity, pharynx, esophagus and larynx, have at greater risk of becoming cancerous than other body tissues. The consumption of alcohol is also associated with an increased risk of stomach, colon, rectum, liver, female breast and ovarian cancers. Conversely, recent studies suggest that red wine components inhibit colony formation of human breast cancer and esophageal carcinoma cells, suggesting that wine-derived phenolic compounds may be inhibitory, in contrast to the alcohol component of wine. Because of a lack of systematic studies dealing with the different types of cancer and alcoholic beverages and wine in particular, in this narrative review we summarize the general risk of cancer linked to the consumption of alcoholic beverages, including wine, according to type of cancer, with 140 extracted relevant references from 1966 to 2020. Mostly epidemiological studies concerning large cohorts have been selected. For the cancers of the upper aerodigestive tract, liver, colorectum, breast cancer, pancreatic, prostate, an excessive consumption and/or misuse of alcoholic beverages is correlated with increased risk. Conversely a probable decreased risk has been found for renal/kidney cancers, as well as for Non-Hodgkin lymphomas, such as thyroid lymphomas, associated with the moderate consumption of alcoholic beverages. There is no evidence of ovarian, gastric, head and neck, and lung cancer being linked to the moderate consumption of alcoholic beverages. Cancer is a multifactorial disease, and many factors contribute to effects on health status, usually being both genetic and environmental. Habits (smoking, dietary/lifestyle pattern/ habits, physical activity), should also be taken into account when defining appropriate consumption frequencies for different types of alcoholic drink (wine, beer, spirits). Further research is needed related to wine consumption in the context of a healthy dietary and lifestyle pattern given health-promoting constituents of wine and its effects on cancer incidence.

Download Full-text

Polynomial Mendelian Randomization reveals widespread non-linear causal effects in the UK Biobank

10.1101/2021.12.08.471751 ◽

2021 ◽

Author(s):

Jonathan Sulc ◽

Jenny Sjaarda ◽

Zoltan Kutalik

Keyword(s):

Causal Inference ◽

Large Scale ◽

Causal Effect ◽

Causal Effects ◽

Mendelian Randomisation ◽

Uk Biobank ◽

Individual Level ◽

Glucose Levels ◽

Causal Function ◽

The Uk

Causal inference is a critical step in improving our understanding of biological processes and Mendelian randomisation (MR) has emerged as one of the foremost methods to efficiently interrogate diverse hypotheses using large-scale, observational data from biobanks. Although many extensions have been developed to address the three core assumptions of MR-based causal inference (relevance, exclusion restriction, and exchangeability), most approaches implicitly assume that any putative causal effect is linear. Here we propose PolyMR, an MR-based method which provides a polynomial approximation of an (arbitrary) causal function between an exposure and an outcome. We show that this method provides accurate inference of the shape and magnitude of causal functions with greater accuracy than existing methods. We applied this method to data from the UK Biobank, testing for effects between anthropometric traits and continuous health-related phenotypes and found most of these (84%) to have causal effects which deviate significantly from linear. These deviations ranged from slight attenuation at the extremes of the exposure distribution, to large changes in the magnitude of the effect across the range of the exposure (e.g. a 1 kg/m2 change in BMI having stronger effects on glucose levels if the initial BMI was higher), to non-monotonic causal relationships (e.g. the effects of BMI on cholesterol forming an inverted U shape). Finally, we show that the linearity assumption of the causal effect may lead to the misinterpretation of health risks at the individual level or heterogeneous effect estimates when using cohorts with differing average exposure levels.

Download Full-text

Body size and composition and site-specific cancers in UK Biobank: a Mendelian randomisation study

10.1101/2020.02.28.970459 ◽

2020 ◽

Cited By ~ 1

Author(s):

Mathew Vithayathil ◽

Paul Carter ◽

Siddhartha Kar ◽

Amy M. Mason ◽

Stephen Burgess ◽

...

Keyword(s):

Instrumental Variables ◽

Association Studies ◽

Genome Wide Association ◽

Mendelian Randomisation ◽

Genome Wide Association Studies ◽

Uk Biobank ◽

Site Specific ◽

Genome Wide ◽

Increased Risk ◽

The Uk

ABSTRACTObjectivesTo investigate the casual role of body mass index, body fat composition and height in cancer.DesignTwo stage mendelian randomisation studySettingPrevious genome wide association studies and the UK BiobankParticipantsGenetic instrumental variables for body mass index (BMI), fat mass index (FMI), fat free mass index (FFMI) and height from previous genome wide association studies and UK Biobank. Cancer outcomes from 367 586 participants of European descent from the UK Biobank.Main outcome measuresOverall cancer risk and 22 site-specific cancers risk for genetic instrumental variables for BMI, FMI, FFMI and height.ResultsGenetically predicted BMI (per 1 kg/m2) was not associated with overall cancer risk (OR 0.99; 95% confidence interval (CI) 0-98-1.00, p=0.105). Elevated BMI was associated with increased risk of stomach cancer (OR 1.15, 95% (CI) 1.05-1.26; p=0.003) and melanoma (OR 0.96, 95% CI 0.92-1.00; p=0.044). For sex-specific cancers, BMI was positively associated with uterine cancer (OR 1.08, 95% CI 1.01-1.14; p=0.015) but inversely associated with breast (OR 0.95, 95% CI 0.92-0.98; p=0.001), prostate (OR 0.95, 95% CI 0.92-0.99; p=0.007) and testicular cancer (OR 0.89, 95% CI 0.81-0.98; p=0.017). Elevated FMI (per 1 kg/m2) was associated with gastrointestinal cancer (stomach cancer OR 4.23, 95% CI 1.18-15.13, p=0.027; colorectal cancer OR 1.94, 95% CI 1.23-3.07; p=0.004). Increased height (per 1 standard deviation, approximately 6.5cm) was associated with increased risk of overall cancer (OR 1.06; 95% 1.04-1.09; p = 2.97×10-8) and most site-specific cancers with the strongest estimates for kidney, non-Hodgkin lymphoma, colorectal, lung, melanoma and breast cancer.ConclusionsThere is little evidence for BMI as a casual risk factor for cancer. BMI may have a causal role for sex-specific cancers, although with inconsistent directions of effect, and FMI for gastrointestinal malignancies. Elevated height is a risk factor for overall cancer and multiple site cancers.

Download Full-text

Association of inflammation with depression and anxiety: evidence for symptom-specificity and potential causality from UK Biobank and NESDA cohorts

Molecular Psychiatry ◽

10.1038/s41380-021-01188-w ◽

2021 ◽

Author(s):

Yuri Milaneschi ◽

Nils Kappelmann ◽

Zheng Ye ◽

Femke Lamers ◽

Sylvain Moser ◽

...

Keyword(s):

Depressive Symptoms ◽

Sleep Problems ◽

Experimental Studies ◽

Anxiety Symptoms ◽

Self Report ◽

Mendelian Randomisation ◽

Depression And Anxiety ◽

Uk Biobank ◽

Increased Risk ◽

The Uk

AbstractWe examined whether inflammation is uniformly associated with all depressive and anxiety symptoms, and whether these associations are potentially causal. Data was from 147,478 individuals from the UK Biobank (UKB) and 2,905 from the Netherlands Study of Depression and Anxiety (NESDA). Circulating C-reactive protein (CRP) was measured in both cohorts and interleukin-6 (IL-6) in NESDA. Genetic instruments for these proteins were obtained from published GWAS and UKB. Depressive and anxiety symptoms were assessed with self-report questionnaires. In NESDA, neurovegetative (appetite, sleep, psychomotor) symptoms were disaggregated as increased vs. decreased. In joint analyses, higher CRP was associated with depressive symptoms of depressed mood (OR = 1.06, 95% CI = 1.05–1.08), altered appetite (OR = 1.25, 95%CI = 1.23–1.28), sleep problems (OR = 1.05, 95%CI = 1.04–1.06), and fatigue (OR = 1.12, 95% CI = 1.11–1.14), and with anxiety symptoms of irritability (OR = 1.06, 95% CI = 1.05–1.08) and worrying control (OR = 1.03, 95% CI = 1.02–1.04). In NESDA, higher IL-6 was additionally associated with anhedonia (OR = 1.30, 95% CI = 1.12–1.52). Higher levels of both CRP (OR = 1.27, 95% CI = 1.13–1.43) and IL-6 (OR = 1.26, 95% CI = 1.07–1.49) were associated with increased sleep. Higher CRP was associated with increased appetite (OR = 1.21, 95% CI = 1.08–1.35) while higher IL-6 with decreased appetite (OR = 1.45, 95% CI = 1.18–1.79). In Mendelian Randomisation analyses, genetically predicted higher IL-6 activity was associated with increased risk of fatigue (estimate = 0.25, SE = 0.08) and sleep problems (estimate = 0.19, SE = 0.07). Inflammation was associated with core depressive symptoms of low mood and anhedonia and somatic/neurovegetative symptoms of fatigue, altered sleep and appetite changes. Less consistent associations were found for anxiety. The IL-6/IL-6R pathway could be causally linked to depression. Experimental studies are required to further evaluate causality, mechanisms, and usefulness of immunotherapies for depressive symptoms.

Download Full-text

Genetic liability to insomnia in relation to cardiovascular diseases: a Mendelian randomisation study

European Journal of Epidemiology ◽

10.1007/s10654-021-00737-5 ◽

2021 ◽

Author(s):

Shuai Yuan ◽

Amy M. Mason ◽

Stephen Burgess ◽

Susanna C. Larsson

Keyword(s):

Cardiovascular Diseases ◽

Association Studies ◽

European Ancestry ◽

Mendelian Randomisation ◽

Genome Wide Association Studies ◽

Uk Biobank ◽

Genetic Liability ◽

Genome Wide ◽

Increased Risk ◽

The Uk

AbstractThe present study aimed to determine the associations between insomnia and cardiovascular diseases (CVDs) using Mendelian randomisation (MR) analysis. As instrumental variables, we used 208 independent single-nucleotide polymorphisms associated with insomnia at the genome-wide significance threshold in a meta-analysis of genome-wide association studies in the UK Biobank and 23andMe including a total of 397 959 self-reported insomnia cases and 933 057 non-cases. Summary-level data for nine CVDs were obtained from the UK Biobank including 367 586 individuals of European ancestry. After correction for multiple testing, genetic liability to insomnia was associated with higher odds of six CVDs, including peripheral arterial disease (odd ratio (OR) 1.22; 95% confidence interval (CI), 1.21, 1.33), heart failure (OR 1.21; 95% CI, 1.13, 1.30), coronary artery disease (OR 1.19; 95% CI, 1.14, 1.25), ischaemic stroke (OR 1.15; 95% CI, 1.06, 1.25), venous thromboembolism (OR 1.13; 95% CI, 1.07, 1.19) and atrial fibrillation (OR 1.10; 95% CI, 1.05, 1.15). There were suggestive associations for aortic valve stenosis (OR, 1.17; 95% CI, 1.04, 1.32) and haemorrhagic stroke (OR 1.14; 95% CI, 1.00, 1.29) but no association for abdominal aortic aneurysm (OR, 1.14, 95% CI, 0.98, 1.33). The patterns of associations remained with mild attenuation in multivariable MR analyses adjusting for genetically correlated phenotypes and potential mediators, including sleep duration, depression, body mass index, type 2 diabetes and smoking. The present MR study suggests potential causal associations of genetic liability to insomnia with increased risk of a broad range of CVDs.

Download Full-text

Smoking and COVID-19 outcomes: an observational and Mendelian randomisation study using the UK Biobank cohort

Thorax ◽

10.1136/thoraxjnl-2021-217080 ◽

2021 ◽

pp. thoraxjnl-2021-217080

Author(s):

Ashley K Clift ◽

Adam von Ende ◽

Pui San Tan ◽

Hannah M Sallis ◽

Nicola Lindson ◽

...

Keyword(s):

Large Scale ◽

Hospital Admissions ◽

Causal Effect ◽

Smoking Status ◽

Smoking Initiation ◽

Mendelian Randomisation ◽

Uk Biobank ◽

Never Smokers ◽

The Uk ◽

Analytical Approaches

BackgroundConflicting evidence has emerged regarding the relevance of smoking on risk of COVID-19 and its severity.MethodsWe undertook large-scale observational and Mendelian randomisation (MR) analyses using UK Biobank. Most recent smoking status was determined from primary care records (70.8%) and UK Biobank questionnaire data (29.2%). COVID-19 outcomes were derived from Public Health England SARS-CoV-2 testing data, hospital admissions data, and death certificates (until 18 August 2020). Logistic regression was used to estimate associations between smoking status and confirmed SARS-CoV-2 infection, COVID-19-related hospitalisation, and COVID-19-related death. Inverse variance-weighted MR analyses using established genetic instruments for smoking initiation and smoking heaviness were undertaken (reported per SD increase).ResultsThere were 421 469 eligible participants, 1649 confirmed infections, 968 COVID-19-related hospitalisations and 444 COVID-19-related deaths. Compared with never-smokers, current smokers had higher risks of hospitalisation (OR 1.80, 95% CI 1.26 to 2.29) and mortality (smoking 1–9/day: OR 2.14, 95% CI 0.87 to 5.24; 10–19/day: OR 5.91, 95% CI 3.66 to 9.54; 20+/day: OR 6.11, 95% CI 3.59 to 10.42). In MR analyses of 281 105 White British participants, genetically predicted propensity to initiate smoking was associated with higher risks of infection (OR 1.45, 95% CI 1.10 to 1.91) and hospitalisation (OR 1.60, 95% CI 1.13 to 2.27). Genetically predicted higher number of cigarettes smoked per day was associated with higher risks of all outcomes (infection OR 2.51, 95% CI 1.20 to 5.24; hospitalisation OR 5.08, 95% CI 2.04 to 12.66; and death OR 10.02, 95% CI 2.53 to 39.72).InterpretationCongruent results from two analytical approaches support a causal effect of smoking on risk of severe COVID-19.

Download Full-text

Influence of blood pressure on pneumonia risk: Epidemiological association and Mendelian randomisation in the UK Biobank

10.1101/2020.04.19.20071936 ◽

2020 ◽

Author(s):

Seyedeh M. Zekavat ◽

Michael Honigberg ◽

James Pirruccello ◽

Puja Kohli ◽

Elizabeth W. Karlson ◽

...

Keyword(s):

Blood Pressure ◽

Observational Studies ◽

Mendelian Randomisation ◽

List Type ◽

Uk Biobank ◽

Increased Risk ◽

Reduce Risk ◽

Novel Coronavirus ◽

The Uk

AbstractObjectivesTo determine whether elevated blood pressure influences risk for respiratory infection.DesignProspective, population-based epidemiological and Mendelian randomisation studies.SettingUK Biobank.Participants377,143 self-identified British descent (54% women; median age 58 years) participants in the UK Biobank.Main outcome measuresFirst incident pneumonia over an average of 8 follow-up years.Results107,310 (30%) participants had hypertension at UK Biobank enrolment, and 9,969 (3%) developed a pneumonia during follow-up. Prevalent hypertension at baseline was significantly associated with increased risk for incident respiratory disease including pneumonia (hazard ratio 1.36 (95% confidence interval 1.29 to 1.43), P<0.001), acute respiratory distress syndrome or respiratory failure (1.43 (1.29 to 1.59), P<0.001), and chronic lower respiratory disease (1.30 (1.25 to 1.36), P<0.001), independent of age, age2, sex, smoking status, BMI, prevalent diabetes mellitus, prevalent coronary artery disease, and principal components of ancestry. Mendelian randomisation analyses indicated that genetic predisposition to a 5 mmHg increase in blood pressure was associated with increased risk of incident pneumonia for SBP (1.08, (1.04 to 1.13), P<0.001) and DBP (1.11 (1.03 to 1.20), P=0.005). Additionally, consistent with epidemiologic associations, increase in blood pressure genetic risk was significantly associated with reduced forced expiratory volume in the first second, forced vital capacity, and the ratio of the two (P<0.001 for all).ConclusionsThese results strongly suggest that elevated blood pressure independently increases risk for pneumonia and reduces pulmonary function. Maintaining adequate blood pressure control, in addition to other measures, may reduce risk for pneumonia. Whether the present findings are generalizable to novel coronavirus disease 2019 (COVID-19) require further study.Summary BoxSection 1: What is already known on this topicHypertension has been associated with pneumonia in small observational studies.Based on early epidemiologic analyses, hypertension is described as a risk factor for SARS-CoV-2 infection and associated novel coronavirus disease 2019 (COVID-19).The influence of hypertension on pneumonia risk is difficult to assess in traditional observational studies.Section 2: What this study addsOur pre-COVID-19 analyses are consistent with a causal relationship between increased blood pressure and increased risk for incident respiratory infections, as well as between increased blood pressure and reduced pulmonary function.These results support hypertension as a pneumonia risk factor; efforts to optimize blood pressure may reduce risk for pneumonia.

Download Full-text

Using genetic variants to evaluate the causal effect of cholesterol lowering on head and neck cancer risk: a Mendelian randomization study

10.1101/2020.10.05.20206268 ◽

2020 ◽

Author(s):

Mark Gormley ◽

James Yarmolinsky ◽

Tom Dudding ◽

Kimberley Burrows ◽

Richard M Martin ◽

...

Keyword(s):

Cancer Risk ◽

Head And Neck ◽

Oropharyngeal Cancer ◽

Causal Effect ◽

Lipid Lowering ◽

List Type ◽

Uk Biobank ◽

Cholesterol Lowering ◽

Increased Risk ◽

Oral And Oropharyngeal Cancer

AbstractIntroductionHead and neck squamous cell carcinoma (HNSCC), which includes cancers of the oral cavity and oropharynx, is a cause of substantial global morbidity and mortality. Strategies to reduce disease burden include discovery of novel therapies and repurposing of existing drugs. Statins are commonly prescribed for lowering circulating cholesterol by inhibiting HMG-CoA reductase (HMGCR). Results from some observational studies suggest that statin use may reduce HNSCC risk. We appraised the relationship of genetically-proxied cholesterol-lowering drug targets and other circulating lipid traits with oral (OC) and oropharyngeal (OPC) cancer risk.Methods and findingsWe conducted two-sample Mendelian randomization (MR). For the primary analysis, germline genetic variants in HMGCR, NPC1L1, CETP, PCSK9 and LDLR were used to proxy the effect of low-density lipoprotein cholesterol (LDL-C) lowering therapies. In secondary analyses, variants were used to proxy circulating levels of other lipid traits in a genome-wide association study (GWAS) meta-analysis of 188,578 individuals. Both primary and secondary analyses aimed to estimate the downstream causal effect of cholesterol lowering therapies on OC and OPC risk.The second sample for MR was taken from a GWAS of 6,034 OC and OPC cases and 6,585 controls (GAME-ON). Analyses were replicated in UK Biobank, using 839 OC and OPC cases and 372,016 controls and the results of the GAME-ON and UK Biobank analyses combined in a fixed-effects meta-analysis.We found limited evidence of a causal effect of genetically-proxied LDL-C lowering using HMGCR, NPC1L1, CETP or other circulating lipid traits on either OC or OPC risk. Genetically-proxied PCSK9 inhibition equivalent to a 1 mmol/L (38.7 mg/dL) reduction in LDL-C was associated with an increased risk of OC and OPC combined (OR 1.8 95%CI 1.2, 2.8, p= 9.31 ×10−05), with good concordance between GAME-ON and UK Biobank (I2= 22%). Effects for PCSK9 appeared stronger in relation to OPC (OR 2.6 95%CI 1.4, 4.9) than OC (OR 1.4 95%CI 0.8, 2.4). LDLR variants, resulting in genetically-proxied reduction in LDL-C equivalent to a 1 mmol/L (38.7 mg/dL), reduced the risk of OC and OPC combined (OR 0.7, 95%CI 0.5, 1.0, p= 0.006). A series of pleiotropy-robust and outlier detection methods showed that pleiotropy did not bias our findings.ConclusionWe found limited evidence for a role of cholesterol-lowering in OC and OPC risk, suggesting previous observational results may have been confounded. There was some evidence that genetically-proxied inhibition of PCSK9 increased risk, while lipid-lowering variants in LDLR, reduced risk of combined OC and OPC. This result suggests that the mechanisms of action of PCSK9 on OC and OPC risk may be independent of its cholesterol lowering effects, but further replication of this finding is required.Author summaryWhy was this study done?To determine if genetically-proxied cholesterol-lowering drugs (such as statins which target HMGCR) reduce oral and oropharyngeal cancer risk.To determine if genetically-proxied circulating lipid traits (e.g. low-density lipoprotein cholesterol) have a causal effect on oral and oropharyngeal cancer risk.What did the researchers do and find?There was little evidence that genetically-proxied inhibition of HMGCR (target of statins), NPC1L1 (target of ezetimibe) and CETP (target of CETP inhibitors) influences oral or oropharyngeal cancer risk.There was little evidence of an effect of circulating lipid traits on oral or oropharyngeal cancer risk.There was some evidence that genetically-proxied inhibition of PCSK9 increases, while lipid-lowering variants in LDLR reduces oral and oropharyngeal cancer risk.What do these findings mean?These findings suggest that the results of previous observational studies examining the effect of statins on oral and oropharyngeal risk may have been confounded.Given we found little evidence of an effect of other cholesterol lowering therapies, the mechanism of action of PCSK9 may be independent of cholesterol-lowering. Further replication of this finding in other head and neck cancer datasets is required.

Download Full-text

Polynomial Mendelian Radomization reveals widespread non-linear causal effects in the UK Biobank

10.21203/rs.3.rs-1152056/v1 ◽

2022 ◽

Author(s):

Jonathan Sulc ◽

Jennifer Sjaarda ◽

Zoltan Kutalik

Keyword(s):

Causal Inference ◽

Large Scale ◽

Causal Effect ◽

Causal Effects ◽

Mendelian Randomisation ◽

Uk Biobank ◽

Individual Level ◽

Glucose Levels ◽

Causal Function ◽

The Uk

Abstract Causal inference is a critical step in improving our understanding of biological processes and Mendelian randomisation (MR) has emerged as one of the foremost methods to efficiently interrogate diverse hypotheses using large-scale, observational data from biobanks. Although many extensions have been developed to address the three core assumptions of MR-based causal inference (relevance, exclusion restriction, and exchangeability), most approaches implicitly assume that any putative causal effect is linear. Here we propose PolyMR, an MR-based method which provides a polynomial approximation of an (arbitrary) causal function between an exposure and an outcome. We show that this method provides accurate inference of the shape and magnitude of causal functions with greater accuracy than existing methods. We applied this method to data from the UK Biobank, testing for effects between anthropometric traits and continuous health-related phenotypes and found most of these (84%) to have causal effects which deviate significantly from linear. These deviations ranged from slight attenuation at the extremes of the exposure distribution, to large changes in the magnitude of the effect across the range of the exposure (e.g. a 1 kg/m2 change in BMI having stronger effects on glucose levels if the initial BMI was higher), to non-monotonic causal relationships (e.g. the effects of BMI on cholesterol forming an inverted U shape). Finally, we show that the linearity assumption of the causal effect may lead to the misinterpretation of health risks at the individual level or heterogeneous effect estimates when using cohorts with differing average exposure levels.

Download Full-text

Examining the association between family status and depression in the UK Biobank

10.1101/2020.07.07.20148023 ◽

2020 ◽

Author(s):

Alexandors Giannelis ◽

Alish Palmos ◽

Saskia P Hagenaars ◽

Gerome Breen ◽

Cathryn M Lewis ◽

...

Keyword(s):

Causal Effect ◽

Age Groups ◽

Risk Scores ◽

Mendelian Randomisation ◽

Family Status ◽

Uk Biobank ◽

Number Of Children ◽

The Uk ◽

Social Demographic ◽

Mental Health Questionnaire

Background: We examined associations between family status (living with a spouse or partner, number of children) and lifetime depression. Methods: We used data from the UK Biobank, a large prospective study of middle-aged and older adults. Lifetime depression was assessed as part of a follow-up mental health questionnaire. Logistic regression was used to estimate associations between family status and depression. We included extensive adjustment for social, demographic and other potential confounders, including depression polygenic risk scores. Results: 52,078 participants (mean age = 63.6, SD = 7.6; 52% female) were included in our analyses. Living with a spouse or partner was associated with substantially lower odds of lifetime depression (OR = 0.67, 95% CI 0.62-0.74). Compared to individuals without children, we found higher odds of lifetime depression for parents of one child (OR = 1.17, 95% CI 1.07-1.27), and parents of three (OR = 1.11, 95% CI 1.03-1.20) or four or more children (OR = 1.27, 95% CI 1.14-1.42). Amongst those not cohabiting, having any number of children was associated with higher odds of lifetime depression. Our results were consistent across age groups, the sexes, neighbourhood deprivation and genetic risk for depression. Exploratory Mendelian randomisation analyses suggested a causal effect of number of children on lifetime depression. Limitations: Our data did not allow distinguishing between non-marital and marital cohabitation. Results may not generalise to all ages or populations. Conclusions: Living with a spouse or partner was strongly associated with reduced odds of depression. Having one or three or more children was associated with increased odds of depression, especially in individuals not living with a spouse or partner.

Download Full-text