CoChat
Sign Up
Cited by 184Genetic Associations and EpidemiologyLipoproteins and Cardiovascular HealthGenetics and Physical Performance

Integrating genomics with biomarkers and therapeutic targets to invigorate cardiovascular drug development

Open access
Holmes, Michael V., Richardson, Tom G., Ference, Brian A., Davies, Neil M., Davey Smith, George·2021·Nature Reviews Cardiology
DOI 10.1038/s41569-020-00493-1
Open this paper in CoChat — free account

Abstract

Drug development in cardiovascular disease is stagnating, with lack of efficacy and adverse effects being barriers to innovation. Human genetics can provide compelling evidence of causation through approaches such as Mendelian randomization, with genetic support for causation increasing the probability of a clinical trial succeeding. Mendelian randomization applied to quantitative traits can identify risk factors for disease that are both causal and amenable to therapeutic modification. However, important differences exist between genetic investigations of a biomarker (such as HDL cholesterol) and a drug target aimed at modifying the same biomarker of interest (such as cholesteryl ester transfer protein), with implications for the methodology, interpretation and application of Mendelian randomization to drug development. Differences include the comparative nature of the genetic architecture - that is, biomarkers are typically polygenic, whereas protein drug targets are influenced by either cis-acting or trans-acting genetic variants - and the potential for drug targets to show disease associations that might differ from those of the biomarker that they are intended to modify (target-mediated pleiotropy). In this Review, we compare and contrast the use of Mendelian randomization to evaluate potential drug targets versus quantitative traits. We explain how genetic epidemiological studies can be used to assess the aetiological roles of biomarkers in disease and to prioritize drug targets, including designing their evaluation in clinical trials.

AI snapshot

Written from the paper’s full text.

TL;DR

This review contrasts Mendelian randomization (MR) applied to complex polygenic biomarkers versus protein drug targets, arguing that they answer different questions and can produce divergent findings. It proposes using biomarker MR for aetiological insight and drug-target MR with cis variants to prioritize targets and inform trial design, while stressing that MR complements rather than replaces randomized controlled trials.

Key findings

  • Biomarker MR asks whether a biomarker causes disease; drug-target MR asks whether modifying a specific drug target alters disease risk. Biomarkers are typically polygenic, whereas drug targets are usually proteins with monogenic or oligogenic genetic architecture.
  • Findings from biomarker MR, drug-target MR, and RCTs can diverge because of target-mediated pleiotropy, off-target drug effects, disease onset versus progression, horizontal pleiotropy, adaptation, canalization, and time-dependent effects. For example, HDL-cholesterol MR provides little causal evidence for CHD, but CETP variants associate with CHD via apolipoprotein B lowering; dalcetrapib raised HDL-C with little/no LDL-C or apoB effect and failed, whereas anacetrapib lowered apoB and showed vascular benefit.
  • Cardiovascular drug development is described as stagnant: approvals were 2.6/year in 2000–2009 and 3.2/year in 2010–2018; phase I launch probability was 4% for cardiovascular drugs versus 9% for anticancer therapies, phase II 10% versus 23%, and phase III 46% versus 59%. Median time from clinical entry to marketing authorization is 10 years, fewer than 1 in 10 drugs succeed from phase I to marketing authorization, and median cost to bring a drug to market in 2009–2018 was approximately $1bn.
  • MR estimates for LDL-C lowering and CHD risk are approximately 3-fold greater than RCT estimates, and the review argues such scaling can help predict trial effects. It also notes that apolipoprotein B appears necessary for lipid-mediated atherogenesis, CRP MR does not support a causal role in CHD, and IL6R genetic evidence supports a causal role consistent with canakinumab trial benefit.

Methods

  • Narrative review, not a primary empirical study; it compares and contrasts MR as applied to quantitative traits versus drug targets, drawing on historical and contemporary articles and illustrative examples.
  • Evidence base consists of published genetic, trial, and drug-development literature, including examples such as CETP, HMGCR, PCSK9, CRP, IL6R, BMI/GLP1R, and sclerostin inhibition.
  • Main analysis is conceptual synthesis: differences in SNP selection (polygenic versus cis-acting), types of pleiotropy, colocalization, PheWAS, scaling of MR estimates to RCT effects, and a proposed framework for identifying therapeutic targets.

Limitations

  • Stated: drug-target MR cannot predict off-target effects of individual therapies; cis-only instruments may lack power and preclude some sensitivity analyses; trans variants are more likely to be pleiotropic; colocalization can suffer from lack of power or suitable data; tissue-specific eQTL/proteomic datasets are modest (GTEx n=21–755); MR estimates reflect lifelong exposure and may exceed short-term RCT effects; MR is not a substitute for clinical trials.
  • Stated: interpretation challenges include time zero, adaptation, canalization, index event bias, and differences between disease onset and progression; GWAS participants were almost 90% European ancestry as of 2017, limiting equity and generalizability.
  • Evident from the text: as a narrative review, it does not conduct a systematic search, meta-analysis, or new quantitative synthesis; examples are selected to illustrate the authors’ arguments.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41569-020-00493-1

Cites

193 of 193 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 131 cited sources checked · 0 supported · 0 flagged

  • Findingcitation
    § When might findings from MR of a complex biomarker, MR of a drug target and findings from an RCT diverge?

    Such SBP associations have not been identified in drug-target MRs using HDL-C raising variants in CETP.

    “such SBP associations have not been identified in drug-target MRs using HDL-C raising variants in CETP 79 77 , for reasons that are unclear.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [79] Benatti, F. et al. Liposuction induces a compensatory increase of visceral fat which is effectively counteracted by physical activity: a randomized trial. J. Clin. Endocrinol. Metab. 97, 2388–2395 (2012). source

      Not checked in the free run.

    • cites [77] Wang, Q. et al. Metabolic profiling of angiopoietin-like protein 3 and 4 inhibition: a drug-target Mendelian randomization analysis. Eur. Heart J. https://doi.org/10.1093/eurheartj/ehaa972 (2020). source

      Not checked in the free run.

  • Findingcitation
    § When might findings from MR of a complex biomarker, MR of a drug target and findings from an RCT diverge?

    In the case of drug-target MR, there is little evidence to support presence of such compensatory mechanisms, e.g. genetic variants with large phenotypic effects appear to have the expected dose-response association with disease as compared to genetic variants of smaller phenotypic effect.

    “in the case of drug-target MR, there is little evidence to support presence of such compensatory mechanisms -e.g. genetic variants with large phenotypic effects appear to have the expected dose-response association with disease as compared to genetic variants of smaller phenotypic effect 57 40 .”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [57] Ridker, P. M. et al. Antiinflammatory therapy with canakinumab for atherosclerotic disease. N. Engl. J. Med. 377, 1119–1131 (2017). source

      Not checked in the free run.

    • cites [40] Burgess, S. et al. Guidelines for performing Mendelian randomization investigations. Wellcome Open Res. 4, 186 (2020). source

      Not checked in the free run.

  • Premisecitation
    § Introduction

    Safety accounts for one quarter of drug development failures.

    “Another major hurdle is safety (accounting for ¼ of failures 6 ),”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [6] Wouters, O. J., McKee, M. & Luyten, J. Estimated research and development investment needed to bring a new medicine to market, 2009–2018. JAMA 323, 844–853 (2020). source

      Not checked in the free run.

  • Premisecitation
    § Please consider and comment on the following points when reviewing this manuscript:

    Human genetics can provide compelling evidence of causation through approaches such as Mendelian randomization.

    “Human genetics can provide compelling evidence of causation through approaches such as Mendelian randomization (MR).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [9] Dowden, H. & Munro, J. Trends in clinical success rates and therapeutic focus. Nat. Rev. Drug Discov. 18, 495–496 (2019). source

      Not checked in the free run.

    • cites [10] Davey Smith, G. & Ebrahim, S. ‘Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease? Int. J. Epidemiol. 32, 1–22 (2003). source

      Not checked in the free run.

  • Premisecitation
    § Introduction

    MR is being widely applied to quantitative traits with the aim of discovering modifiable risk factors for disease that are causal and amenable to therapeutic modification.

    “With increasing availability of large-scale phenotyping (and genotyping), MR is being widely applied to quantitative traits (e.g. markers or adiposity, such as body mass index 13-16 ; blood lipids such as HDL-cholesterol 17-19 ; inflammation-related proteins such as interleukin-6 receptor 20,21 12 ) with the aim of discovering modifiable risk factors for disease that are causal and amenable to the”
    Quote in the paperSubstantial claim0 of 9 sources checked
    • cites [13] Davies, N. M., Holmes, M. V. & Davey Smith, G. Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians. BMJ 362, k601 (2018). source

      Not checked in the free run.

    • cites [14] Nelson, M. R. et al. The support of human genetic evidence for approved drug indications. Nat. Genet. 47, 856–860 (2015). source

      Not checked in the free run.

    • cites [15] Zheng, J. et al. Phenome-wide Mendelian randomization mapping the influence of the plasma proteome on complex diseases. Nat. Genet. 52, 1122–1131 (2020). source

      Not checked in the free run.

    • cites [16] Lyall, D. M. et al. Association of body mass index with cardiometabolic disease in the UK biobank: a mendelian randomization study. JAMA Cardiol. 1, 882–889 (2017). source

      Not checked in the free run.

    • cites [17] Dale, C. E. et al. Causal associations of adiposity and body fat distribution with coronary heart disease, stroke subtypes, and type 2 diabetes mellitus: a mendelian randomization analysis. Circulation 135, 2373–2388 (2017). source

      Not checked in the free run.

    • cites [18] Emdin, C. A. et al. Genetic association of waist-to-hip ratio with cardiometabolic traits, type 2 diabetes, and coronary heart disease. JAMA 317, 626–634 (2017). source

      Not checked in the free run.

    • cites [19] Holmes, M. V. et al. Causal effects of body mass index on cardiometabolic traits and events: a Mendelian randomization analysis. Am. J. Hum. Genet. 94, 198–208 (2014). source

      Not checked in the free run.

    • cites [20] Voight, B. F. et al. Plasma HDL cholesterol and risk of myocardial infarction: a mendelian randomisation study. Lancet 380, 572–580 (2012). source

      Not checked in the free run.

    • cites [21] White, J. et al. Association of lipid fractions with risks for coronary artery disease and diabetes. JAMA Cardiol. 1, 692–699 (2016). source

      Not checked in the free run.

  • Classified as probably not a substantial claim. 60
  • May combine two claims. 45

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given
    § Please consider and comment on the following points when reviewing this manuscript:

    Drug targets can show pleiotropic disease associations that may differ from those of the biomarker they are designed to modify.

    “and the potential for drug targets to show pleiotropic disease associations which may differ to those of the biomarker it is designed to modify.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation
    § Introduction

    Fewer than 1 in 10 drugs succeed from phase I through to marketing authorization, with the success rate being disease-specific.

    “It is largely for these reasons that fewer than 1 in 10 drugs succeed in running the gauntlet from phase I through to marketing authorization, with the success rate being disease-specific 7 .”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [7] Harrison, R. K. Phase II and phase III failures: 2013–2015. Nat. Rev. Drug Discov. 15, 817–818 (2016). source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § Introduction

    For drugs entering phase II RCTs, the probability of launch was 10% for cardiovascular indications and 23% for anticancer indications.

    “For drugs entering phase II RCTs the comparable probabilities of launch were 10% for those with cardiovascular indications and 23% for anticancer indications”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [8] Alteri, E. & Guizzaro, L. Be open about drug failures to speed up research. Nature 563, 317–319 (2018). source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § Introduction

    The average number of new drug approvals for cardiovascular indications was 2.6 per year from 2000-2009 and 3.2 per year from 2010-2018.

    “the corresponding values for cardiovascular indications were 2.6/year and 3.2/year 2 .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [2] WHO. Cardiovascular diseases (CVDs). https://www.who.int/news-room/fact-sheets/detail/cardiovascular-diseases-(cvds) (2017).

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § Why focus on biomarkers?

    Adiposity increases systolic blood pressure, which mediates part of the effect of BMI on risk of coronary artery disease.

    “A further example is the causal role that adiposity plays in CHD: adiposity increases systolic blood pressure 13,16 which mediates part of the effect of BMI on risk of CAD 55 .”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [13] Davies, N. M., Holmes, M. V. & Davey Smith, G. Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians. BMJ 362, k601 (2018). source

      Not checked in the free run.

    • cites [16] Lyall, D. M. et al. Association of body mass index with cardiometabolic disease in the UK biobank: a mendelian randomization study. JAMA Cardiol. 1, 882–889 (2017). source

      Not checked in the free run.

    • cites [55] Clinton, S. K. & Libby, P. Cytokines and growth factors in atherogenesis. Arch. Pathol. Lab. Med. 116, 1292–1300 (1992).

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given
    § Biomarker versus drug target

    Most drug targets are proteins.

    “Most drug targets are proteins.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation
    § Why focus on biomarkers?

    Pharmacological modification of C-reactive protein is unlikely to be beneficial to coronary heart disease, whereas randomized evidence demonstrates the value of targeting biomarkers further upstream, for example interleukin-6 receptor or interleukin 1-beta.

    “as pharmacological modification of C-reactive protein is unlikely to be beneficial to CHD 51 , whereas randomized evidence demonstrates the value of targeting biomarkers further upstream, e.g. interleukin-6 receptor 21,54 or interleukin 1-beta 52”
    Quote in the paperSubstantial claim0 of 4 sources checked
    • cites [51] Swanson, S. A., Tiemeier, H., Ikram, A. I. & Hernán, M. A. Nature as a trialist? Deconstructing the analogy between Mendelian randomization and randomized trials. Epidemiology 28, 653–659 (2017). source

      Not checked in the free run.

    • cites [21] White, J. et al. Association of lipid fractions with risks for coronary artery disease and diabetes. JAMA Cardiol. 1, 692–699 (2016). source

      Not checked in the free run.

    • cites [54] Libby, P. Inflammation in atherosclerosis. Nature 420, 868–874 (2002). source

      Not checked in the free run.

    • cites [52] Ference, B. A., Holmes, M. V. & Davey Smith, G. Using Mendelian randomization to improve the design of randomized trials. Cold Spring Harb. Perspect. Med. https://doi.org/10.1101/cshperspect.a040980 (2021). source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisenone given
    § Biomarker versus drug target

    There is an important distinction between a complex polygenic biomarker and an oligogenic drug target.

    “There is an important distinction between a complex (polygenic) biomarker and an (oligogenic) drug target, which we outline in Figure 1 .”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Premisenone given
    § Biomarker versus drug target

    Complex traits have polygenic architectures, with multiple genetic variants affecting the biomarker through discrete and/or overlapping pathways.

    “Complex traits have polygenic architectures, with multiple genetic variants affecting the biomarker through discrete and/or overlapping pathways.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisereasoning
    § Biomarker versus drug target

    If the therapeutic aim is to treat or prevent disease through pharmacological modification of a complex biomarker, this is typically conducted by exploiting the vertically pleiotropic effects of the protein on the complex biomarker.

    “Thus, if the therapeutic aim is to treat or prevent disease through the pharmacological modification of a complex biomarker, this is typically conducted by exploiting the vertically pleiotropic effects of the protein on the complex biomarker”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation
    § Why focus on biomarkers?

    Blood lipid traits including apolipoprotein B, low-density lipoprotein cholesterol and triglycerides each associate independently with risk of coronary heart disease.

    “For example, blood lipid traits (including apolipoprotein B, low-density lipoprotein cholesterol and triglycerides) each associate independently with risk of CHD.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [57] Ridker, P. M. et al. Antiinflammatory therapy with canakinumab for atherosclerotic disease. N. Engl. J. Med. 377, 1119–1131 (2017). source

      Not checked in the free run.

    • cites [58] Tardif, J. C. et al. Efficacy and safety of low-dose colchicine after myocardial infarction. N. Engl. J. Med. 381, 2497–2505 (2019). source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisereasoning
    § Mendelian randomization to elucidate causality

    Genetic variants are inherited in offspring from their parents at random, owing to Mendel's law of independent assortment.

    “The process by which genetic variants are inherited in offspring from their parents occurs at random , owing to Mendel's law of independent assortment.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisereasoning
    § Biomarker versus drug target

    Trans variants may associate with protein levels at conventional GWAS significance thresholds but are more likely to have horizontally pleiotropic effects, making the interpretation of MR estimates based on trans-acting variants challenging.

    “While genetic variants genomically distant from the protein-encoding gene (so-called trans variants) may associate with protein levels at conventional GWAS significance thresholds, they are more likely to have horizontally pleiotropic effects (e.g. by affecting the expression of multiple genes), making the interpretation of MR estimates based on trans-acting variants challenging ( Figure 6 ).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisereasoning
    § Biomarker versus drug target

    A genetic instrument utilising cis-acting variants for a drug target is typically monogenic but can sometimes be oligogenic if more than one region of the genome is involved in encoding the final protein product.

    “Thus a genetic instrument utilising cis-acting variants for a drug target is typically monogenic but can sometimes be oligogenic if more than one region of the genome (as in SNPs from more than one locus) is involved in encoding the final protein product.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning
    § Biomarker versus drug target

    A drug-target MR of haemoglobin selecting cis-acting variants from across discrete protein-encoding genes would be oligogenic, in contrast to a monogenic instrument for CRP, encoded for by a single locus.

    “Thus a drug-target MR of haemoglobin selecting cis-acting variants from across discrete protein-encoding genes would be oligogenic, in contrast to a monogenic instrument for CRP 51 , encoded for by a single locus.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [51] Swanson, S. A., Tiemeier, H., Ikram, A. I. & Hernán, M. A. Nature as a trialist? Deconstructing the analogy between Mendelian randomization and randomized trials. Epidemiology 28, 653–659 (2017). source

      Not checked in the free run.

  • May combine two claims.

    Premisenone given
    § Biomarker versus drug target

    Proteins are typically encoded for by a unique single gene, making it theoretically possible to generate a genetic instrument for the protein of interest that is specific, so-called cis protein instruments.

    “Proteins are typically encoded for by a unique (single) gene making it theoretically possible to generate a genetic instrument for the protein of interest that is specific (socalled 'cis' protein instruments that are close to, or in, the relevant gene, see Box 6 ).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given
    § Biomarker versus drug target

    If more than one SNP is used as an instrument from the same locus, this remains a monogenic instrument.

    “Notably, if more than one SNP is used as an instrument from the same locus, this remains a monogenic instrument.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Premisenone given
    § What does this mean in practice?

    BMI is a polygenic trait, and a genetic instrument for BMI consists of many SNPs that associate with BMI through multiple mechanisms.

    “BMI is a polygenic trait, and a genetic instrument for BMI consists of many SNPs that associate with BMI through multiple mechanisms.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation
    § What does this mean in practice?

    CETP inhibitors and HMGCR both lower coronary heart disease through pharmacological reductions of apolipoprotein B, but their metabolomic signatures differ wildly.

    “At a more extreme level, while both CETP inhibitors and HMGCR lower CHD through the pharmacological reductions of apolipoprotein B 22 , their metabolomic signatures differ wildly 68 .”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [22] Holmes, M. V. et al. Mendelian randomization of blood lipids for coronary heart disease. Eur. Heart J. 36, 539–550 (2015). source

      Not checked in the free run.

    • cites [68] Rose, G. Sick individuals and sick populations. Int. J. Epidemiol. 14, 32–38 (1985). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § Biomarker versus drug target

    Haemoglobin is a heterotetramer that is made up of alpha and beta globin polypeptides, each encoded by discrete genes.

    “For example, haemoglobin is a heterotetramer that is made up of alpha and beta globin polypeptides, each encoded by discrete genes 62 .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [62] Ference, B. A. et al. Association of triglyceride-lowering LPL variants and LDL-C-lowering LDLR variants with risk of coronary heart disease. JAMA 321, 364–373 (2019). source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning
    § What does this mean in practice?

    A drug target may have different effects as compared to a biomarker, and different drugs that modify the same biomarker may differ in their effects on risk factors and disease outcomes.

    “Thus, a drug target may have different effects as compared to a biomarker, and different drugs that modify the same biomarker may differ in their effects on risk factors and disease outcomes.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationreasoning
    § What does this mean in practice?

    Environmental changes in BMI, for example through cosmetic liposuction or weight loss surgery, might lead to effects on cardiovascular risk factors and disease that are specific to each surgical procedure.

    “Environmental changes in BMI (e.g. through cosmetic liposuction or weight loss surgery) might lead to effects on cardiovascular risk factors and disease that are specific to each surgical procedure.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § When might findings from MR of a complex biomarker, MR of a drug target and findings from an RCT diverge?

    Individuals at higher genetic predisposition to obesity smoke more, which could represent an attempt to self-administer smoking for its nicotinic effects on appetite suppression and weight loss.

    “individuals at higher genetic predisposition to obesity who smoke more 87,88 , which could represent an attempt to self-administer smoking for its nicotinic effects on appetite suppression and weight loss 89 .”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [87] Sofat, R. et al. Separating the mechanism-based and off-target actions of cholesteryl ester transfer protein inhibitors with CETP gene polymorphisms. Circulation 121, 52–62 (2010). source

      Not checked in the free run.

    • cites [88] Dzeshka, M. S., Shahid, F., Shantsila, A. & Lip, G. Y. H. Hypertension and atrial fibrillation: an intimate association of epidemiology, pathophysiology, and outcomes. Am. J. Hypertens. 30, 733–755 (2017). source

      Not checked in the free run.

    • cites [89] January, C. T. et al. 2019 AHA/ACC/HRS focused update of the 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society. Circulation 140, e125–e151 (2019). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisereasoning
    § When might findings from MR of a complex biomarker, MR of a drug target and findings from an RCT diverge?

    Local linkage disequilibrium can distort findings.

    “it is plausible that local linkage disequilibrium can distort findings.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation
    § When might findings from MR of a complex biomarker, MR of a drug target and findings from an RCT diverge?

    All CETP inhibitors show modest SBP increasing effects, and this appears to be most pronounced for torcetrapib.

    “while all CETP inhibitors show modest SBP increasing effects, this appears to be most pronounced for torcetrapib 78 74 .”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [78] Husain, M. et al. Oral semaglutide and cardiovascular outcomes in patients with type 2 diabetes. N. Engl. J. Med. 381, 841–851 (2019). source

      Not checked in the free run.

    • cites [74] Yarmolinsky, J. et al. Association between genetically proxied inhibition of HMG-CoA reductase and epithelial ovarian cancer. JAMA 323, 646–655 (2020). source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § When might findings from MR of a complex biomarker, MR of a drug target and findings from an RCT diverge?

    Colocalization can be used to test for such bias, but can suffer from issues of lack of power and/or availability of suitable data from relevant tissues.

    “Colocalization, which is a statistical approach that quantifies the extent to which two traits share the same underlying causal variant, and thus can be used to infer vertical pleiotropy 83-85 , can be used to test for such bias 86 , but can suffer from issues of lack of power and/or availability of suitable data from relevant tissues.”
    Quote in the paperNot a substantial claim0 of 4 sources checked
    • cites [83] Armitage, J., Holmes, M. V. & Preiss, D. Cholesteryl ester transfer protein inhibition for preventing cardiovascular events. J. Am. Coll. Cardiol. 73, 477–487 (2019). source

      Not checked in the free run.

    • cites [84] HPS3/TIMI55-REVEAL Collaborative Group. Effects of anacetrapib in patients with atherosclerotic vascular disease. N. Engl. J. Med. 377, 1217–1227 (2017). source

      Not checked in the free run.

    • cites [85] Schwartz, G. G. et al. Effects of dalcetrapib in patients with a recent acute coronary syndrome. N. Engl. J. Med. 367, 2089–2099 (2012). source

      Not checked in the free run.

    • cites [86] Barter, P. J. et al. Effects of torcetrapib in patients at high risk for coronary events. N. Engl. J. Med. 357, 2109–2212 (2007). source

      Not checked in the free run.

  • May combine two claims.

    Premisereasoning
    § When might findings from MR of a complex biomarker, MR of a drug target and findings from an RCT diverge?

    A biomarker may be non-causal for disease, but the drug target may show clinical efficacy.

    “A biomarker may be non-causal for disease, but the drug target may show clinical efficacy (scenario 3 in Table 2 ).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § When might findings from MR of a complex biomarker, MR of a drug target and findings from an RCT diverge?

    Methods are under development for applying MR to disease progression, which take into account issues such as index event bias.

    “Methods are under development for applying MR to disease progression, 80 which take into account issues such as index event bias 81 .”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [80] Mohammed, B. S., Cohen, S., Reeds, D., Young, V. L. & Klein, S. Long-term effects of large-volume liposuction on metabolic risk factors for coronary heart disease. Obesity 16, 2648–2651 (2008). source

      Not checked in the free run.

    • cites [81] Bovijn, J. et al. Evaluating the cardiovascular safety of sclerostin inhibition using evidence from meta-analysis of clinical trials and human genetics. Sci. Transl. Med. 12, eaay6570 (2020). source

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation
    § What does this mean in practice?

    Comparing the effects on coronary heart disease of modifying bone mineral density implied by genetic variants in the SOST gene to the effects implied by other BMD associated genetic variants across the genome found very similar disease associations, suggesting that BMD increasing therapies in general, including sclerostin inhibitors, are likely to increase risk of coronary heart disease rather než

    “Very similar disease associations were found suggesting that BMD increasing therapies in general, including sclerostin inhibitors, are likely to increase risk of CHD (rather than representing a case of target-mediated pleiotropy, i.e. an effect specific to sclerostin inhibition; see Box 5 for definition).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [73] Meikle, P. J. et al. Statin action favors normalization of the plasma lipidome in the atherogenic mixed dyslipidemia of MetS: potential relevance to statin-associated dysglycemia. J. Lipid Res. 56, 2381–2392 (2015). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given
    § Putting the pieces together

    The use of human genetics can facilitate drug development throughout the lifecycle.

    “the use of human genetics can facilitate drug development throughout the lifecycle.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisereasoning
    § Conclusions

    The motivation for biomarker and drug target MR differs, with biomarkers informing the effects of potential public health measures and shedding light on causal risk factors for disease that motivate their pharmacological modification, while drug target MRs estimate the likelihood of a specific therapeutic showing efficacy in a clinical trial.

    “The motivation for biomarker and drug target MR differs, with biomarkers informing the effects of potential public health measures and shedding light on causal risk factors for disease that motivate their pharmacological modification, while drug target MRs estimate the likelihood of a specific therapeutic (typically designed to modify a complex biomarker) showing efficacy in a clinical trial.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given
    § Key points

    Proteins are modified into complex biomarkers.

    “Proteins, encoded by genetic variants, are modified into complex biomarkers.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given
    § Conclusions

    A biomarker is typically a complex phenotype that is optimally instrumented through MR using multiple genetic variants across the genome.

    “A biomarker is typically a complex phenotype that is optimally instrumented through MR using multiple genetic variants across the genome.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation
    § When might findings from MR of a complex biomarker, MR of a drug target and findings from an RCT diverge?

    The approach outlined here to obtaining a scaling estimate has generally held true for lipid-lowering therapies and risk of cardiovascular disease, but there are several potential reasons why the derivation of such a scaling for other therapeutic target-indication pairs may pose a challenge.

    “The approach outlined here to obtaining a scaling estimate has generally held true for lipid-lowering therapies and risk of cardiovascular disease 22,23,40,92 but there are several potential reasons why the derivation of such a scaling for other therapeutic target-indication pairs may pose a challenge.”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites [22] Holmes, M. V. et al. Mendelian randomization of blood lipids for coronary heart disease. Eur. Heart J. 36, 539–550 (2015). source

      Not checked in the free run.

    • cites [23] Interleukin-6 Receptor Mendelian Randomisation Analysis Consortium. The interleukin-6 receptor as a target for prevention of coronary heart disease: a mendelian randomisation analysis. Lancet 379, 1214–1224 (2012). source

      Not checked in the free run.

    • cites [40] Burgess, S. et al. Guidelines for performing Mendelian randomization investigations. Wellcome Open Res. 4, 186 (2020). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given
    § Key points

    Proteins are encoded by genetic variants.

    “Proteins, encoded by genetic variants, are modified into complex biomarkers.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning
    § Conclusions

    In both scenarios, human genetics can be harnessed as an invaluable tool to elucidate causal roles in disease aetiology and through careful application, such studies can be used to guide the design and interpretation of clinical trials of medicines.

    “In both scenarios, human genetics can be harnessed as an invaluable tool to elucidate causal roles in disease aetiology and through careful application, such studies can be used to guide the design and interpretation of clinical trials of medicines.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation
    § When might findings from MR of a complex biomarker, MR of a drug target and findings from an RCT diverge?

    Approaches such as estimating the causal effects of a biomarker at different points in life using multivariable MR may help provide evidence about timing, but this approach has not been used for drug targets and may not be possible owing to the limited repertoire of genetic instruments.

    “Approaches such as estimating the causal effects of a biomarker at different points in life using multivariable MR may help provide evidence about timing (as has been conducted for BMI 91 ), however this approach has not been used for drug targets, and may not be possible owing to the limited repertoire of (particularly cis-acting) genetic instruments.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [91] Mahmoud, O., Dudbridge, F., Smith, G. D., Munafo, M. & Tilling, K. Slope-Hunter: a robust method for index-event bias correction in genome-wide association studies of subsequent traits. bioRxiv https://doi.org/10.1101/2020.01.31.928077 (2020). source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisenone given
    § Conclusions

    Drug target MR typically focuses on a protein trait using cis-acting genetic variants in an individual gene that encodes the protein.

    “drug target MR typically focuses on a protein trait using cis -acting genetic variants in an individual gene that encodes the protein.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given
    § Key points

    A biomarker is typically a complex trait that has a polygenic architecture.

    “A biomarker is typically a complex trait that has a polygenic architecture, whereas a drug target is usually a protein that has a monogenic architecture.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning
    § Key points

    These differences have implications for how each source of evidence contributes to aetiological insights in disease and in predicting the effects of therapies on clinical outcomes.

    “These differences have implications for how each source of evidence contributes to aetiological insights in disease and in predicting the effects of therapies on clinical outcomes”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given
    § Key points

    Some proteins and complex biomarkers play a causal role in cardiovascular disease.

    “Some proteins and complex biomarkers play a causal role in cardiovascular disease.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning
    § Putting the pieces together

    MR studies of these proteins for CHD would characterize, for each protein, whether it is likely to be on the causal pathway from therapeutic target to disease.

    “MR studies 497 of these proteins for CHD would characterize, for each protein, whether it is likely to be on the 498 causal pathway from therapeutic target to disease.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given
    § Key points

    Mendelian randomization offers unique opportunities to explore the likely causal role of biomarkers and drug targets in disease aetiology.

    “Mendelian randomization offers unique opportunities to explore the likely causal role of biomarkers and drug targets in disease aetiology.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given
    § Putting the pieces together

    If a biomarker is identified to have a particular relationship with disease (e.g. long latency or time dependency), then individual drug targets that modify the same biomarker ought to display the same attributes in their relationship with disease.

    “If a biomarker is identified to have a particular relationship with disease (e.g. long 488 latency or time dependency), then individual drug targets that modify the same biomarker 489 ought to display the same attributes in their relationship with disease.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation
    § Putting the pieces together

    Colchicine and icosapent ethyl have both demonstrated efficacy in reduction of CVD in recent large-scale phase III trials, however the precise mechanisms by which these drugs act, i.e. the causal intermediaries, remains elusive.

    “colchicine 53 and icosapent ethyl 97 have both demonstrated efficacy in reduction of CVD in 494 recent large-scale phase III trials, however the precise mechanisms by which these drugs act, 495 i.e. the causal intermediaries, remains elusive.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [53] Strimbu, K. & Tavel, J. A. What are biomarkers? Curr. Opin. HIV AIDS 5, 463–466 (2010). source

      Not checked in the free run.

    • cites [97] Landmesser, U. et al. 2017 Update of ESC/EAS Task Force on practical clinical guidance for proprotein convertase subtilisin/kexin type 9 inhibition in patients with atherosclerotic cardiovascular disease or in familial hypercholesterolaemia. Eur. Heart J. 39, 1131–1143 (2017). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given
    § Key points

    A drug target is usually a protein that has a monogenic architecture.

    “A biomarker is typically a complex trait that has a polygenic architecture, whereas a drug target is usually a protein that has a monogenic architecture.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning
    § Putting the pieces together

    Data from these various sources can be used to triangulate evidence and predict the likely outcome of a therapeutic trial.

    “Data from these various sources can be used to triangulate evidence and predict the likely outcome of a therapeutic trial.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Premisenone given
    § Key points

    Biomarker MR motivation is public health and biological understanding to discover biomarkers to intervene on.

    “Public health (e.g. implications of obesity on disease) Biological understanding to discover biomarkers to intervene on”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisenone given
    § Key points

    In contrast, a drug-target MR typically uses one or a handful of SNPs selected from the gene that encodes an individual protein (cis-acting variants) in a monogenic instrument.

    “In contrast, a drug-target MR typically uses one or a handful of SNPs selected from the gene that encodes an individual protein ( cis -acting variants) in a monogenic instrument.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given
    § Key points

    It may be possible to proceed from disease GWAS through to drug-target MR without elucidating causal risk factors, but this is likely to have less power to do so.

    “Of note, it may be possible to proceed 1075 from disease GWAS through to drug-target MR without elucidating causal risk factors, but this 1076 is likely to have less power to do so.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisenone given
    § Key points

    For complex biomarker MR, a genetic instrument typically consists of multiple SNPs selected from across the genome (a polygenic instrument) associated with the biomarker of interest.

    “For complex biomarker MR, a genetic instrument typically consists of multiple SNPs selected from across the genome (a polygenic instrument) associated with the biomarker of interest.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning
    § Key points

    Such may lead to the development of new therapies that stem the tide of harmful exposures (e.g. excess adiposity) that are proving challenging to modify at a population level.

    “Such may lead to the development of new therapies that stem the tide of harmful exposures (e.g. excess adiposity) that are proving challenging to modify at a population level.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationreasoning
    § Key points

    The distinction between a protein trait (being typically monogenic) and a complex biomarker (being polygenic) is critical to differentiating the motivation, conduct, interpretation and application of MR to drug targets (i.e. proteins) as compared to complex biomarkers.

    “It is this distinction that is critical to differentiating the motivation, conduct, interpretation and application of MR to drug targets (i.e. proteins) as compared to complex biomarkers.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationreasoning
    § Key points

    Care needs to be placed in interpreting the magnitudes of effect given that MR of an exposure represents, naively, lifetime exposure and depending on the exposure-outcome pair, this may lead to a magnification of the effect estimates arising from MR.

    “As discussed in the text, care needs to be 1072 placed in interpreting the magnitudes of effect given that MR of an exposure represents, 1073 naively, lifetime exposure and depending on the exposure-outcome pair, this may lead to a 1074 magnification of the effect estimates arising from MR.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationcitation
    § Key points

    An exploration through available resources such as OpenTargets or ChEMBL can provide additional information to establish whether individual loci are likely to represent tractable therapeutic targets.

    “An exploration 1067 through available resources such as OpenTargets or ChEMBL 105 can provide additional 1068 information to establish whether individual loci are likely to represent tractable therapeutic 1069 targets.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [105] Silverwood, R. J. et al. Testing for non-linear causal effects using a binary genotype in a Mendelian randomization study: application to alcohol and cardiovascular traits. Int. J. Epidemiol. 43, 1781–1790 (2014). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given
    § Key points

    Drug target MR asks whether modifying a drug target alters risk of disease.

    “Does modifying this drug target alter of disease?”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given
    § Key points

    Biomarker MR asks whether a biomarker causes disease.

    “Does this biomarker cause disease?”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning
    § Key points

    Drug-target MR of protein-encoding loci identified as being of relevance to risk factors can establish the predicted effects of pharmacological modification in clinical trials and be used to guide clinical development.

    “Drug-target MR of protein-encoding loci identified as being of relevance to risk factors 1070 can establish the predicted effects of pharmacological modification in clinical trials and be used 1071 to guide clinical development (as shown in Figure 2 ).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given
    § Key points

    Drug target MR motivation is new therapies for prevention of incident and recurrent disease, and disease-related sequelae.

    “New therapies for the prevention of incident and recurrent disease, and disease-related sequelae”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationnone given
    § 1082

    Scenario 2: Biomarker doesn't cause disease, nor does the drug target.

    “Biomarker doesn't cause disease, nor does the drug target.”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: Studies are adequately powered.

  • May combine two claims.

    Premisecitation
    § Key points

    Vertical pleiotropy can shed light on differences in treatments (e.g. PCSK9 and HMGCR) on intermediaries (e.g. blood metabolic markers), which can then be used to explore whether such differences are clinically relevant.

    “Can shed light on differences in treatments (e.g. PCSK9 and HMGCR) on intermediaries (e.g. blood metabolic markers 106 ), which can then be used to explore whether such differences are clinically”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [106] Lawlor, D. A., Harbord, R. M., Sterne, J. A., Timpson, N. & Davey Smith, G. Mendelian randomization: using genes as instruments for making causal inferences in epidemiology. Stat. Med. 27, 1133–1163 (2008). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given
    § Key points

    Limitations for drug target MR: cannot predict off-target effects of individual therapies.

    “Cannot predict off-target effects of individual therapies”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Premisenone given
    § Key points

    Role of PheWAS for drug target MR is unknown target-mediated effects (both unintended: ADRs, and beneficial: repurposing).

    “Unknown (both unintended: ADRs, and beneficial: repurposing) target- mediated effects”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Premisenone given
    § Key points

    Drug target MR might be less vulnerable to pleiotropy but should ideally check consistency of independent SNPs in same locus.

    “Might be less vulnerable to pleiotropy but should ideally check consistency of independent SNPs in same locus”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Premisenone given
    § Key points

    Role of PheWAS for biomarker MR is unknown effects of biomarker on other biomarkers and diseases.

    “Unknown effects biomarker on other biomarkers and diseases”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given
    § Key points

    Limitations for biomarker MR: public health interventions may not reciprocate effects.

    “Public health interventions may not reciprocate effects (as challenges in purely”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given
    § Key points

    For biomarker MR, SNP selection is GWAS-derived preferable (unless protein exposure).

    “GWAS-derived preferable (unless protein exposure)”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given
    § Key points

    Causal inference for biomarker MR is evidence on whether the biomarker causes incident disease and progression of disease.

    “Evidence on whether the biomarker causes incident disease and progression of disease”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationnone given
    § 1082

    Scenario 1: Biomarker causes disease; and modifying biomarker through the drug target has same effect on disease.

    “Biomarker causes disease; and modifying biomarker through the drug target has same effect on disease.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: Studies are adequately powered.

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given
    § Key points

    Causal inference for drug target MR is evidence on whether modifying a drug target is likely to alter risk of disease onset (general population) or disease progression (established disease).

    “Evidence on whether modifying a drug target is likely to alter risk of (i) disease onset (when the study is set in general population); or (ii) disease progression (when the study is set in individuals with established disease)”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation
    § 1082

    REVEAL trial of anacetrapib demonstrated evidence of vascular benefit; RCTs of apoA-I infusion have thus far not shown benefit.

    “REVEAL trial of anacetrapib 75 demonstrated evidence of vascular benefit; RCTs of apoA-I infusion 111 have thus far not shown benefit”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [75] Sliz, E. et al. Metabolomic consequences of genetic inhibition of PCSK9 compared with statin treatment. Circulation 138, 2499–2512 (2018). source

      Not checked in the free run.

    • cites [111] Deaton, A. & Cartwright, N. Understanding and misunderstanding randomized controlled trials. Soc. Sci. Med. 210, 2–21 (2018). source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § 1082

    BMI increases risk of T2D and CHD in MR, with one potential mechanism being increased blood glucose.

    “BMI increases risk of T2D and CHD in MR 14 with one potential mechanism being increased blood glucose”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [14] Nelson, M. R. et al. The support of human genetic evidence for approved drug indications. Nat. Genet. 47, 856–860 (2015). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given
    § 1082

    Scenario 3: Drug target shows target-mediated pleiotropy that leads to beneficial effect on disease.

    “Drug target shows target-mediated pleiotropy that leads to beneficial effect on disease.”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: Studies are adequately powered.

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given
    § Key points

    Vertical pleiotropy can point to pharmacological approaches to intervention (e.g. proteomic mediators of adiposity and BMI/CAD).

    “Can point to pharmacological approaches to intervention (e.g. proteomic mediators of adiposity and BMI/CAD)”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given
    § Key points

    Biomarker MR methods are heavily reliant on sensitivity analyses to show lack of unbalanced horizontal pleiotropy.

    “Heavily reliant on sensitivity analyses to show lack of unbalanced horizontal pleiotropy”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation
    § 1082

    CETP variants associate with risk of CHD; APOA1 variants do not associate with risk of CHD.

    “CETP 23 variants associate with of CHD; APOA1 117 variants do not associate with risk of CHD”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [23] Interleukin-6 Receptor Mendelian Randomisation Analysis Consortium. The interleukin-6 receptor as a target for prevention of coronary heart disease: a mendelian randomisation analysis. Lancet 379, 1214–1224 (2012). source

      Not checked in the free run.

    • cites [117] Ference, B. A., Majeed, F., Penumetcha, R., Flack, J. M. & Brook, R. D. Effect of Naturally random allocation to lower low-density lipoprotein cholesterol on the risk of coronary heart disease mediated by polymorphisms in NPC1L1, HMGCR, or both. J. Am. Coll. Cardiol. 65, 1552–1561 (2015). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given
    § Key points

    For drug target MR, cis-acting SNPs are preferable.

    “Cis -acting SNPs where possible”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationcitation
    § Key points

    Horizontal pleiotropy needs to be formally evaluated, e.g. through robust MR approaches.

    “Need to formally evaluate, e.g. through robust MR approaches 37”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [37] Brumpton, B. et al. Avoiding dynastic, assortative mating, and population stratification biases in Mendelian randomization through within-family analyses. Nat. Commun. 11, 3519 (2020). source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § Box 9. Promoting equitable development of therapeutic targets through Mendelian randomization

    Women and minorities have historically been under investigated and under-represented in clinical research.

    “Women and minorities have historically been under investigated and under-represented in clinical research 153 .”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [153] Holmes, M. V., Perel, P., Shah, T., Hingorani, A. D. & Casas, J. P. CYP2C19 genotype, clopidogrel metabolism, platelet function, and cardiovascular events: a systematic review and meta-analysis. JAMA 306, 2704–2714 (2011). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationnone given
    § Box 4: How to conduct the optimal MR for a biomarker vs a drug target? 1120

    The advantages of using cis SNPs include specificity, and the possibility of assessing such specificity through co-localisation.

    “The advantages of using cis SNPs include specificity, and the possibility of assessing such specificity through co-localisation.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § Box 8. Heterogeneity of treatment effects

    A factorial study of MR predicted effects of PCSK9 inhibition and HMGCR inhibition on CHD and type II diabetes exists.

    “as, for example, in a factorial study of MR predicted effects of PCSK9 inhibition and HMGCR inhibition on CHD and type II diabetes 107 .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [107] Ference, B. A. Using genetic variants in the targets of lipid lowering therapies to inform drug discovery and development: current and future treatment options. Clin. Pharmacol. Ther. 105, 568–581 (2019). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 1082

    Phospholipase A2 (PLA2) is a family of enzymes that hydrolyse fatty acid from membrane phospholipids.

    “Phospholipase A2 (PLA 2 ) is a family of enzymes that hydrolyse fatty acid from membrane phospholipids 125 .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [125] Look Ahead Research Group. Cardiovascular effects of intensive lifestyle intervention in type 2 diabetes. N. Engl. J. Med. 369, 145–154 (2013). source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § 1082

    The interleukin 6 receptor (IL6R) is a drug target that alters CRP, and genetic evidence supports a causal role of IL6R in CHD.

    “The interleukin 6 receptor (IL6R) is a drug target that alters CRP (see Figure 3 ), and for which genetic evidence supports a causal role of IL6R in CHD 20 54”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [20] Voight, B. F. et al. Plasma HDL cholesterol and risk of myocardial infarction: a mendelian randomisation study. Lancet 380, 572–580 (2012). source

      Not checked in the free run.

    • cites [54] Libby, P. Inflammation in atherosclerosis. Nature 420, 868–874 (2002). source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given
    § Box 2: How can genetic evidence can enhance the evidence base of causation to facilitate the 1090 conduct of randomized controlled trials? 1091

    Drug target Mendelian randomization studies can provide a source of valuable information that is of value to understanding whether the modification of a drug target is likely to lead to clinical benefit.

    “Drug target Mendelian randomization studies can provide a source of valuable information that is of value to understanding whether the modification of a drug target is likely to lead to clinical benefit.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisenone given
    § 1082

    Secretory phospholipase A2-IIA (sPLA2-IIA) and lipoprotein-associated PLA2 (Lp-PLA2) were both thought to play potential roles in atherogenesis.

    “Secretory phospholipase A2-IIA (sPLA 2 -IIA) and lipoprotein-associated PLA2 (Lp-PLA 2 ) were both thought to play potential roles in atherogenesis.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationreasoning
    § Box 4: How to conduct the optimal MR for a biomarker vs a drug target? 1120

    Use of trans SNPs may lead to biased estimates but can potentially be compared to the cis instrument for comparability and analysed for pleiotropy using well-established MR approaches.

    “Use of trans SNPs by definition includes SNPs that associate through either an upstream mechanism or gene regulatory functions, that may lead to biased estimates (see Figure 6 ), but which can potentially be (a) compared to the cis instrument for comparability and (b) analysed for pleiotropy using well-established MR approaches.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning
    § Box 8. Heterogeneity of treatment effects

    These analyses can be useful for prospectively designing clinical trials, in order to maximize effect estimates, and also as an approach to stratified medicines.

    “These analyses can be useful for prospectively designing clinical trials, in order to maximize effect estimates, and also as an approach to stratified medicines.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationnone given
    § Box 4: How to conduct the optimal MR for a biomarker vs a drug target? 1120

    The disadvantages of using cis-only include potential lack of power owing to an insufficient proportion of variance of the exposure explained, and potential inability to conduct sensitivity analyses that require multiple variants.

    “The disadvantages of using cisonly includes potential lack of power owing to an insufficient proportion of variance of the exposure explained, and potential inability to conduct sensitivity analyses that require multiple variants (e.g. MR-Egger and weighted median approaches, inter alia ).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning
    § Box 6: Cis and trans genetic variants

    Cis-acting variants are likely to have effects specific to the protein of interest, and are less likely to be pleiotropic.

    “By altering the exposure of interest (i.e. the protein) through the gene that directly encodes it, such cis-acting variants are likely to have effects specific to the protein of interest, and are less likely to be pleiotropic.”
    Quote in the paperSubstantial claim
89 of 89

  • The paper relies on this without support.

    Most MR studies use either case control studies where the case has a first event and controls are disease free, or prospective cohorts of the general population with individuals followed up for incident events.

    “Most MR studies use either case control studies where the case has a first event and controls are disease free, or prospective cohorts of the general population with individuals followed up for incident events.”

  • The paper relies on this without support.

    In both circumstances, the SNP to disease association captures the risk of incident disease.

    “In both circumstances, the SNP to disease association captures the risk of incident disease.”

  • The paper relies on this without support.

    Many trials are set in patients either at high risk of CVD or patients who have in the past experienced a CHD event.

    “many trials are set in patients either at high risk of CVD (who likely have prevalent subclinical atherosclerosis at entry into the trial) or patients who have in the past experienced a CHD event.”

  • The paper relies on this without support.

    In CVD, risk factors for onset (e.g. blood pressure, lipids) appear to be similar for first and recurrent disease.

    “Although in CVD, risk factors for onset (e.g. blood pressure, lipids) appear to be similar for first and recurrent disease”

  • The paper relies on this without support.

    A cause of AF may be hypertension, however once AF is established, its persistence is likely to be influenced to a greater extent by drugs involved in controlling the rate and rhythm, as opposed to drugs treating the initiating cause.

    “a cause of AF may be hypertension, however once AF is established, its persistence is likely to be influenced to a greater extent by drugs involved in controlling the rate and rhythm, as opposed to drugs treating the initiating cause (e.g. lowering blood pressure).”

  • The paper relies on this without support.

    The timing of causal effects for each exposure-disease pair is likely to differ.

    “The timing of causal effects for each exposure-disease pair is likely to differ.”

  • The authors state this as an assumption.

    Studies are adequately powered.

    “potential explanations assume that studies are adequately powered.”
7 of 7

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Using multiple genetic instruments in combination improves statistical power because the combination increases the total fraction of the exposure variance explained by the instruments.

    “Using multiple genetic instruments in combination improves statistical power because the combination increases the total fraction of the exposure variance explained by the instruments.27,41”

    From Mendelian randomization (2022) · cites this paper as [41]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Conflicting competing-interest statements for GDS

    Minor

    GDS is among the authors not otherwise named in the competing-interest declarations, so the statement that all other authors have no conflicts conflicts with the immediately following disclosure of research support from GSK and Biogen.

1 of 1

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Holmes, M. V., Richardson, T. G., Ference, B. A., Davies, N. M., & Davey Smith, G.
Affiliations
  • University of Oxford (GB)
  • University of Bristol (GB)
  • University of Cambridge (GB)
  • Norwegian University of Science and Technology (NO)
Corresponding author
Holmes, Michael V.
ORCID
Venue
Nature Reviews Cardiology
Publisher
Springer Science and Business Media LLC
Published
2021-03-11
Volume
18
Issue
6
Pages
435-453
ISSN
  • 1759-5002 (ISSN-L)
  • 1759-5010
DOI
10.1038/s41569-020-00493-1
OpenAlex
W3080477490
Cited by
184
References
193
Citations by year
2026: 21 · 2025: 46 · 2024: 45 · 2023: 32 · 2022: 26 · 2021: 13 · 2020: 1
FWCI
4.38
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Reviews Cardiology (publishedVersion)
  • PubMed (publishedVersion)
  • Bristol Research (University of Bristol) (submittedVersion, other-oa)
Funders
  • Cancer Research UK (16896)
  • British Heart Foundation (FS/18/23/33512)
  • Medical Research Council (MC_UU_00011/1, MR/S003886/1)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genetic Associations and Epidemiology
  • Health Sciences › Medicine › Surgery › Lipoproteins and Cardiovascular Health
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genetics and Physical Performance
Keywords
cardiovascular drug development, Mendelian randomization, biomarkers, therapeutic target, genetic epidemiology, cholesteryl ester transfer protein, HDL cholesterol, causal inference, clinical trial design, cardiovascular disease
MeSH terms
Drug Development, Cardiovascular Agents, Cardiovascular Diseases, Humans, Biomarkers, Genomics

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-10.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-10.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review