Principled distillation of UK Biobank phenotype data reveals underlying structure in human variation.

Data within biobanks capture broad yet detailed indices of human variation, but biobank-wide insights can be difficult to extract due to complexity and scale. Here, using large-scale factor analysis, we distill hundreds of variables (diagnoses, assessments and survey items) into 35 latent constructs, using data from unrelated individuals with predominantly estimated European genetic ancestry in UK Biobank. These factors recapitulate known disease classifications, disentangle elements of socioeconomic status, highlight the relevance of psychiatric constructs to health and improve measurement of pro-health behaviours. We go on to demonstrate the power of this approach to clarify genetic signal, enhance discovery and identify associations between underlying phenotypic structure and health outcomes. In building a deeper understanding of ways in which constructs such as socioeconomic status, trauma, or physical activity are structured in the dataset, we emphasize the importance of considering the interwoven nature of the human phenome when evaluating public health patterns.

Implications of the genomic revolution for education research and policy.

Research at the intersection of social science and genomics, 'sociogenomics', is transforming our understanding of the interplay between genomics, individual outcomes and society. It has interesting and maybe unexpected implications for education research and policy. Here we review the growing sociogenomics literature and discuss its implications for educational researchers and policymakers. We cover key concepts and methods in genomic research into educational outcomes, how genomic data can be used to investigate social or environmental effects, the methodological strengths and limitations of genomic data relative to other observational social data, the role of intergenerational transmission and potential policy implications. The increasing availability of genomic data in studies can produce a wealth of new evidence for education research. This may provide opportunities for disentangling the environmental and genomic factors that influence educational outcomes and identifying potential mechanisms for intervention.

Challenging the current hypothesis that thrombosis is responsible for the post-COVID-19 condition.

People with the post-COVID-19 condition suffer symptoms that persist beyond 12 weeks following acute COVID-19 infection. Fatigue, shortness of breath, and cognitive dysfunction ("brain fog") are common. Scientists, clinicians, and patients debate the pathophysiology. One pathophysiological hypothesis is that prothrombotic changes associated with acute COVID-19 persist, causing clots that lead to symptoms. This theory, arising from a research team in South Africa and supported by a paper in Nature Medicine, has been widely disseminated on social media and entered the public narrative as a cause of the post-COVID-19 condition. We describe the development of this theory, examine the findings of a Cochrane review that critically appraises the "microclot" beliefs, and critically appraise the influential study relating clotting biomarkers to cognitive deficits. We conclude the inferences for the hypothesis are not based on evidence, unlicensed use of antithrombotic medication is not justified, and apheresis should not be considered outside of a well-designed clinical trial.

Familial confounding or measurement error? How to interpret findings from sibling and co-twin control studies.

Epidemiological researchers often examine associations between risk factors and health outcomes in non-experimental designs. Observed associations may be causal or confounded by unmeasured factors. Sibling and co-twin control studies account for familial confounding by comparing exposure levels among siblings (or twins). If the exposure-outcome association is causal, the siblings should also differ regarding the outcome. However, such studies may sometimes introduce more bias than they alleviate. Measurement error in the exposure may bias results and lead to erroneous conclusions that truly causal exposure-outcome associations are confounded by familial factors. The current study used Monte Carlo simulations to examine bias due to measurement error in sibling control models when the observed exposure-outcome association is truly causal. The results showed that decreasing exposure reliability and increasing sibling-correlations in the exposure led to deflated exposure-outcome associations and inflated associations between the family mean of the exposure and the outcome. The risk of falsely concluding that causal associations were confounded was high in many situations. For example, when exposure reliability was 0.7 and the observed sibling-correlation was r = 0.4, about 30-90% of the samples (n = 2,000) provided results supporting a false conclusion of confounding, depending on how p-values were interpreted as evidence for a family effect on the outcome. The current results have practical importance for epidemiological researchers conducting or reviewing sibling and co-twin control studies and may improve our understanding of observed associations between risk factors and health outcomes. We have developed an app (SibSim) providing simulations of many situations not presented in this paper.

Childhood adiposity underlies numerous adult brain traits commonly attributed to midlife obesity.

Obese adults are often reported to have smaller brain volumes than their non-obese peers. Whether this represents evidence of accelerations in obesity-driven atrophy or is instead a legacy of developmental differences established earlier in the lifespan remains unclear. This study aimed to investigate whether early-life differences in adiposity explain differences in numerous adult brain traits commonly attributed to mid-life obesity. We utilised a two-sample lifecourse Mendelian randomization study in 37,501 adults recruited to UK Biobank (UKB) imaging centers from 2014, with secondary analyses in 6,996 children assessed in the Adolescent Brain Cognitive Development Study (ABCD) recruited from 2018. Exposures were genetic variants for childhood (266 variants) and adult (470 variants) adiposity derived from a GWAS of 407,741 UKB participants. Primary outcomes were adult total brain volume; grey matter volume, thickness, and surface area; white matter volume and hyperintensities; and hippocampus, amygdala, and thalamus volumes at mean age 55 in UKB. Secondary outcomes were equivalent childhood measures collected at mean age 10 in ABCD. In UKB, individuals who were genetically-predicted to have had higher levels of adiposity in childhood were found to have multiple smaller adult brain volumes relative to intracranial volume (e.g. z-score difference in normalised brain volume per category increase in adiposity [95%CI] = -0.20 [-0.28, -0.12]; p = 4 × 10-6). These effect sizes remained essentially unchanged after accounting for birthweight or current adult obesity in multivariable models, whereas most observed adult effects attenuated towards null (e.g. adult z-score [95%CI] for total volume = 0.06 [-0.05,0.17]; p = 0.3). Observational analyses in ABCD showed a similar pattern of changes already present in those with a high BMI by age 10 (z-score [95%CI] = -0.10 [-0.13, -0.07]; p = 8 × 10-13), with follow-up genetic risk score analyses providing some evidence for a causal effect already at this early age. Sensitivity analyses revealed that many of these effects were likely due to the persistence of larger head sizes established in those who gained excess weight in childhood (childhood z-score [95%CI] for intracranial volume = 0.14 [0.05,0.23]; p = 0.002), rather than smaller brain sizes per se. Our data suggest that persistence of early-life developmental differences across the lifecourse may underlie numerous neuroimaging traits commonly attributed to obesity-related atrophy in later life.

Non-linear Mendelian randomization: detection of biases using negative controls with a focus on BMI, Vitamin D and LDL cholesterol.

Mendelian randomisation (MR) is an established technique in epidemiological investigation, using the principle of random allocation of genetic variants at conception to estimate the causal linear effect of an exposure on an outcome. Extensions to this technique include non-linear approaches that allow for differential effects of the exposure on the outcome depending on the level of the exposure. A widely used non-linear method is the residual approach, which estimates the causal effect within different strata of the non-genetically predicted exposure (i.e. the "residual" exposure). These "local" causal estimates are then used to make inferences about non-linear effects. Recent work has identified that this method can lead to estimates that are seriously biased, and a new method-the doubly-ranked method-has been introduced as a possibly more robust approach. In this paper, we perform negative control outcome analyses in the MR context. These are analyses with outcomes onto which the exposure should have no predicted causal effect. Using both methods we find clearly biased estimates in certain situations. We additionally examined a situation for which there are robust randomised controlled trial estimates of effects-that of low-density lipoprotein cholesterol (LDL-C) reduction onto myocardial infarction, where randomised trials have provided strong evidence of the shape of the relationship. The doubly-ranked method did not identify the same shape as the trial data, and for LDL-C and other lipids they generated some highly implausible findings. Therefore, we suggest there should be extensive simulation and empirical methodological examination of performance of both methods for NLMR under different conditions before further use of these methods. In the interim, use of NLMR methods needs justification, and a number of sanity checks (such as analysis of negative and positive control outcomes, sensitivity analyses excluding removal of strata at the extremes of the distribution, examination of biological plausibility and triangulation of results) should be performed.

Mammographic density mediates the protective effect of early-life body size on breast cancer risk.

The unexplained protective effect of childhood adiposity on breast cancer risk may be mediated via mammographic density (MD). Here, we investigate a complex relationship between adiposity in childhood and adulthood, puberty onset, MD phenotypes (dense area (DA), non-dense area (NDA), percent density (PD)), and their effects on breast cancer. We use Mendelian randomization (MR) and multivariable MR to estimate the total and direct effects of adiposity and age at menarche on MD phenotypes. Childhood adiposity has a decreasing effect on DA, while adulthood adiposity increases NDA. Later menarche increases DA/PD, but when accounting for childhood adiposity, this effect is attenuated. Next, we examine the effect of MD on breast cancer risk. DA/PD have a risk-increasing effect on breast cancer across all subtypes. The MD SNPs estimates are heterogeneous, and additional analyses suggest that different mechanisms may be linking MD and breast cancer. Finally, we evaluate the role of MD in the protective effect of childhood adiposity on breast cancer. Mediation MR analysis shows that 56% (95% CIs [32%-79%]) of this effect is mediated via DA. Our finding suggests that higher childhood adiposity decreases mammographic DA, subsequently reducing breast cancer risk. Understanding this mechanism is important for identifying potential intervention targets.

Ancestry, ethnicity, and race: explaining inequalities in cardiometabolic disease.

Population differences in cardiometabolic disease remain unexplained. Misleading assumptions over genetic explanations are partly due to terminology used to distinguish populations, specifically ancestry, race, and ethnicity. These terms differentially implicate environmental and biological causal pathways, which should inform their use. Genetic variation alone accounts for a limited fraction of population differences in cardiometabolic disease. Research effort should focus on societally driven, lifelong environmental determinants of population differences in disease. Rather than pursuing population stratifiers to personalize medicine, we advocate removing socioeconomic barriers to receipt of and adherence to healthcare interventions, which will have markedly greater impact on improving cardiometabolic outcomes. This requires multidisciplinary collaboration and public and policymaker engagement to address inequalities driven by society rather than biology per se.

Differences between brachial and aortic blood pressure in adolescence and their implications for diagnosis of hypertension.

OBJECTIVES: Blood pressure (BP) is the leading global cause of mortality, and its prevalence is increasing in children and adolescents. Aortic BP is lower than brachial BP in adults. We aimed to assess the extent of this difference and its impact on the diagnosis of hypertension among adolescents. METHODS: We used data from 3850 participants from a UK cohort of births in the early 1990s in the Southwest of England, who attended their ∼17-year follow-up and had valid measures of brachial and aortic BP at that clinic [mean (SD) age 17.8 (0.4) years, 66% female individuals]. Data are presented as mean differences [95% prediction intervals] for both sexes. RESULTS: Aortic systolic BP (SBP) was lower than brachial SBP [male, -22.3 (-31.2, -13.3) mmHg; female, -17.8 (-25.5, -10.0) mmHg]. Differences between aortic and brachial diastolic BP (DBP) were minimal. Based on brachial BP measurements, 101 male individuals (6%) and 22 female individuals (1%) were classified as hypertensive. In contrast, only nine male individuals (<1%) and 14 female individuals (<1%) met the criteria for hypertension based on aortic BP, and the predictive value of brachial BP for aortic hypertension was poor (positive-predictive value = 13.8%). Participants with aortic hypertension had a higher left ventricular mass index than those with brachial hypertension. CONCLUSION: Brachial BP substantially overestimates aortic BP in adolescents because of marked aortic-to-brachial pulse pressure amplification. The use of brachial BP measurement may result in an overdiagnosis of hypertension during screening in adolescence.

Assessing causal links between age at menarche and adolescent mental health: a Mendelian randomisation study.

BACKGROUND: The timing of puberty may have an important impact on adolescent mental health. In particular, earlier age at menarche has been associated with elevated rates of depression in adolescents. Previous research suggests that this relationship may be causal, but replication and an investigation of whether this effect extends to other mental health domains is warranted. METHODS: In this Registered Report, we triangulated evidence from different causal inference methods using a new wave of data (N = 13,398) from the Norwegian Mother, Father, and Child Cohort Study. We combined multiple regression, one- and two-sample Mendelian randomisation (MR), and negative control analyses (using pre-pubertal symptoms as outcomes) to assess the causal links between age at menarche and different domains of adolescent mental health. RESULTS: Our results supported the hypothesis that earlier age at menarche is associated with elevated depressive symptoms in early adolescence based on multiple regression (ß = - 0.11, 95% CI [- 0.12, - 0.09], pone-tailed < 0.01). One-sample MR analyses suggested that this relationship may be causal (ß = - 0.07, 95% CI [- 0.13, 0.00], pone-tailed = 0.03), but the effect was small, corresponding to just a 0.06 standard deviation increase in depressive symptoms with each earlier year of menarche. There was also some evidence of a causal relationship with depression diagnoses during adolescence based on one-sample MR (OR = 0.74, 95% CI [0.54, 1.01], pone-tailed = 0.03), corresponding to a 29% increase in the odds of receiving a depression diagnosis with each earlier year of menarche. Negative control and two-sample MR sensitivity analyses were broadly consistent with this pattern of results. Multivariable MR analyses accounting for the genetic overlap between age at menarche and childhood body size provided some evidence of confounding. Meanwhile, we found little consistent evidence of effects on other domains of mental health after accounting for co-occurring depression and other confounding. CONCLUSIONS: We found evidence that age at menarche affected diagnoses of adolescent depression, but not other domains of mental health. Our findings suggest that earlier age at menarche is linked to problems in specific domains rather than adolescent mental health in general.

Mendelian Randomization analysis of the causal effect of cigarette smoking on hospital costs.

INTRODUCTION: Knowledge of the impact of smoking on healthcare costs is important for establishing the external effects of smoking and for evaluating policies intended to modify this behavior. Conventional analysis of this association is difficult because of omitted variable bias, reverse causality, and measurement error. METHODS: We approached these challenges using a Mendelian Randomization study design; genetic variants associated with smoking behaviors were used in instrumental variables models with inpatient hospital costs (calculated from electronic health records) as the outcome. We undertook genome wide association studies to identify genetic variants associated with smoking initiation and a composite smoking index (reflecting cumulative health impacts of smoking) on up to 300,045 individuals (mean age: 57 years at baseline, range 39 to 72 years) in the UK Biobank. We followed individuals up for a mean of six years. RESULTS: Genetic liability to initiate smoking (ever versus never smoking) was estimated to increase mean per-patient annual inpatient hospital costs by £477 (95% confidence interval (CI): £187 to £766). A one-unit change in genetic liability to the composite smoking index (range: 0-4.0) increased inpatient hospital costs by £204 (95% CI: £105 to £303) per unit increase in this index. There was some evidence that the composite smoking index causal models violated the instrumental variable assumptions, and all Mendelian Randomization models were estimated with considerable uncertainty. Models conditioning on risk tolerance were not robust to weak instrument bias. CONCLUSIONS: Our findings have implications for the potential cost-effectiveness of smoking interventions. IMPLICATIONS: We report the first Mendelian Randomization analysis of the causal effect of smoking on healthcare costs. Using two distinct smoking phenotypes, we identified substantial impacts of smoking on inpatient hospital costs, although the causal models were associated with considerable uncertainty. These results could be used alongside other evidence on the impact of smoking to evaluate the cost-effectiveness of anti-smoking interventions and to understand the scale of externalities associated with this behaviour.

Prostate-Specific Antigen Screening and 15-Year Prostate Cancer Mortality: A Secondary Analysis of the CAP Randomized Clinical Trial.

Importance: The Cluster Randomized Trial of PSA Testing for Prostate Cancer (CAP) reported no effect of prostate-specific antigen (PSA) screening on prostate cancer mortality at a median 10-year follow-up (primary outcome), but the long-term effects of PSA screening on prostate cancer mortality remain unclear. Objective: To evaluate the effect of a single invitation for PSA screening on prostate cancer-specific mortality at a median 15-year follow-up compared with no invitation for screening. Design, Setting, and Participants: This secondary analysis of the CAP randomized clinical trial included men aged 50 to 69 years identified at 573 primary care practices in England and Wales. Primary care practices were randomized between September 25, 2001, and August 24, 2007, and men were enrolled between January 8, 2002, and January 20, 2009. Follow-up was completed on March 31, 2021. Intervention: Men received a single invitation for a PSA screening test with subsequent diagnostic tests if the PSA level was 3.0 ng/mL or higher. The control group received standard practice (no invitation). Main Outcomes and Measures: The primary outcome was reported previously. Of 8 prespecified secondary outcomes, results of 4 were reported previously. The 4 remaining prespecified secondary outcomes at 15-year follow-up were prostate cancer-specific mortality, all-cause mortality, and prostate cancer stage and Gleason grade at diagnosis. Results: Of 415 357 eligible men (mean [SD] age, 59.0 [5.6] years), 98% were included in these analyses. Overall, 12 013 and 12 958 men with a prostate cancer diagnosis were in the intervention and control groups, respectively (15-year cumulative risk, 7.08% [95% CI, 6.95%-7.21%] and 6.94% [95% CI, 6.82%-7.06%], respectively). At a median 15-year follow-up, 1199 men in the intervention group (0.69% [95% CI, 0.65%-0.73%]) and 1451 men in the control group (0.78% [95% CI, 0.73%-0.82%]) died of prostate cancer (rate ratio [RR], 0.92 [95% CI, 0.85-0.99]; P = .03). Compared with the control, the PSA screening intervention increased detection of low-grade (Gleason score [GS] ≤6: 2.2% vs 1.6%; P < .001) and localized (T1/T2: 3.6% vs 3.1%; P < .001) disease but not intermediate (GS of 7), high-grade (GS ≥8), locally advanced (T3), or distally advanced (T4/N1/M1) tumors. There were 45 084 all-cause deaths in the intervention group (23.2% [95% CI, 23.0%-23.4%]) and 50 336 deaths in the control group (23.3% [95% CI, 23.1%-23.5%]) (RR, 0.97 [95% CI, 0.94-1.01]; P = .11). Eight of the prostate cancer deaths in the intervention group (0.7%) and 7 deaths in the control group (0.5%) were related to a diagnostic biopsy or prostate cancer treatment. Conclusions and Relevance: In this secondary analysis of a randomized clinical trial, a single invitation for PSA screening compared with standard practice without routine screening reduced prostate cancer deaths at a median follow-up of 15 years. However, the absolute reduction in deaths was small. Trial Registration: isrctn.org Identifier: ISRCTN92187251.

Characterizing the Causal Pathway From Childhood Adiposity to Right Heart Physiology and Pulmonary Circulation Using Lifecourse Mendelian Randomization.

BACKGROUND: Observational epidemiological studies have reported an association between childhood adiposity and altered cardiac morphology and function in later life. However, whether this is due to a direct consequence of being overweight during childhood has been difficult to establish, particularly as accounting for other measures of body composition throughout the lifecourse can be exceptionally challenging. METHODS AND RESULTS: In this study, we used human genetics to investigate this using a causal inference technique known as lifecourse Mendelian randomization. This approach allowed us to evaluate the effect of childhood body size on 11 measures of right heart and pulmonary circulation independent of other anthropometric traits at various stages in the lifecourse. We found strong evidence that childhood body size has a direct effect on an enlarged right heart structure in later life (eg, right ventricular end-diastolic volume: ß=0.24 [95% CI, 0.15-0.33]; P=3×10-7) independent of adulthood body size. In contrast, childhood body size effects on maximum ascending aorta diameter attenuated upon accounting for body size in adulthood, suggesting that this effect is likely attributed to individuals remaining overweight into later life. Effects of childhood body size on pulmonary artery traits and measures of right atrial function became weaker upon accounting for adulthood fat-free mass and childhood height, respectively. CONCLUSIONS: Our findings suggest that, although childhood body size has a long-term influence on an enlarged heart structure in adulthood, associations with the other structural components of the cardiovascular system and their function may be largely attributed to body composition at other stages in the lifecourse.

Impact of vaccination on the association of COVID-19 with cardiovascular diseases: An OpenSAFELY cohort study.

Infection with SARS-CoV-2 is associated with an increased risk of arterial and venous thrombotic events, but the implications of vaccination for this increased risk are uncertain. With the approval of NHS England, we quantified associations between COVID-19 diagnosis and cardiovascular diseases in different vaccination and variant eras using linked electronic health records for ~40% of the English population. We defined a 'pre-vaccination' cohort (18,210,937 people) in the wild-type/Alpha variant eras (January 2020-June 2021), and 'vaccinated' and 'unvaccinated' cohorts (13,572,399 and 3,161,485 people respectively) in the Delta variant era (June-December 2021). We showed that the incidence of each arterial thrombotic, venous thrombotic and other cardiovascular outcomes was substantially elevated during weeks 1-4 after COVID-19, compared with before or without COVID-19, but less markedly elevated in time periods beyond week 4. Hazard ratios were higher after hospitalised than non-hospitalised COVID-19 and higher in the pre-vaccination and unvaccinated cohorts than the vaccinated cohort. COVID-19 vaccination reduces the risk of cardiovascular events after COVID-19 infection. People who had COVID-19 before or without being vaccinated are at higher risk of cardiovascular events for at least two years.

PheWAS-based clustering of Mendelian Randomisation instruments reveals distinct mechanism-specific causal effects between obesity and educational attainment.

Mendelian Randomisation (MR) estimates causal effects between risk factors and complex outcomes using genetic instruments. Pleiotropy, heritable confounders, and heterogeneous causal effects violate MR assumptions and can lead to biases. To alleviate these, we propose an approach employing a Phenome-Wide association Clustering of the MR instruments (PWC-MR) and apply this method to revisit the surprisingly large apparent causal effect of body mass index (BMI) on educational attainment (EDU): [Formula: see text] = -0.19 [-0.22, -0.16]. First, we cluster 324 BMI-associated genetic instruments based on their association with 407 traits in the UK Biobank, which yields six distinct groups. Subsequent cluster-specific MR reveals heterogeneous causal effect estimates on EDU. A cluster enriched for socio-economic indicators yields the largest BMI-on-EDU causal effect estimate ([Formula: see text] = -0.49 [-0.56, -0.42]) whereas a cluster enriched for body-mass specific traits provides a more likely estimate ([Formula: see text] = -0.09 [-0.13, -0.05]). Follow-up analyses confirms these findings: within-sibling MR ([Formula: see text] = -0.05 [-0.09, -0.01]); MR for childhood BMI on EDU ([Formula: see text] = -0.03 [-0.06, -0.002]); step-wise multivariable MR ([Formula: see text] = -0.05 [-0.07, -0.02]) where socio-economic indicators are jointly modelled. Here we show how the in-depth examination of the BMI-EDU causal relationship demonstrates the utility of our PWC-MR approach in revealing distinct pleiotropic pathways and confounder mechanisms.

Using the global randomization test as a Mendelian randomization falsification test for the exclusion restriction assumption.

Mendelian randomization may give biased causal estimates if the instrument affects the outcome not solely via the exposure of interest (violating the exclusion restriction assumption). We demonstrate use of a global randomization test as a falsification test for the exclusion restriction assumption. Using simulations, we explored the statistical power of the randomization test to detect an association between a genetic instrument and a covariate set due to (a) selection bias or (b) horizontal pleiotropy, compared to three approaches examining associations with individual covariates: (i) Bonferroni correction for the number of covariates, (ii) correction for the effective number of independent covariates, and (iii) an r2 permutation-based approach. We conducted proof-of-principle analyses in UK Biobank, using CRP as the exposure and coronary heart disease (CHD) as the outcome. In simulations, power of the randomization test was higher than the other approaches for detecting selection bias when the correlation between the covariates was low (r2 < 0.1), and at least as powerful as the other approaches across all simulated horizontal pleiotropy scenarios. In our applied example, we found strong evidence of selection bias using all approaches (e.g., global randomization test p < 0.002). We identified 51 of the 58 CRP genetic variants as horizontally pleiotropic, and estimated effects of CRP on CHD attenuated somewhat to the null when excluding these from the genetic risk score (OR = 0.96 [95% CI: 0.92, 1.00] versus 0.97 [95% CI: 0.90, 1.05] per 1-unit higher log CRP levels). The global randomization test can be a useful addition to the MR researcher's toolkit.