The causal association of bilirubin with cardiovascular disease: Are there still any questions?

Searching for genes predisposing to complex diseases improved dramatically by genome-wide association studies (GWAS) 15 years ago. Before, it was similar a search needle in the haystack. Thirty ago, one had have first some evidence that at least “major gene” is influencing phenotype of interest. This came often from twin or segregation analysis done large pedigrees with Nowadays, these methods faded into background since GWAS are applied each independent on whether there any priori genetic influence certain not. agnostic unbiased approach revealed many surprises during last decade. For bilirubin, started “ancient” times. Hunt and colleagues analyses found “major” gene elevating serum bilirubin concentrations Utah [[1]Hunt S.C. Wu L.L. Hopkins P.N. Williams R.R. Evidence major concentration pedigrees.Arterioscler. Thromb. Vasc. Biol. 1996; 16: 912-917Crossref PubMed Scopus (40) Google Scholar,[2]Hunt Kronenberg F. Eckfeldt J.H. Myers R.H. Heiss G. Association plasma coronary heart disease as trait: NHLBI Family Heart Study.Atherosclerosis. 2001; 154: 747-754Abstract Full Text PDF (75) Scholar]. paved way linkage identify responsible chromosomal region 2q34-37 most promising [[3]Kronenberg Coon H. Gutin A. Abkevich V. Samuels M.E. Ballinger D.G. A genome scan loci anti-atherogenic levels.Eur. J. Hum. Genet. 2002; 10: 539-546Crossref (36) Scholar], which has been confirmed [[4]Lin J.P. Cupples L.A. Wilson P.W. Heard-Costa N. O'Donnell C.J. chromosome 2q telomere: genomewide Framingham study.Am. 2003; 72: 1029-1034Abstract (46) Scholar,[5]Lin J.-P. Schwaiger Zheng Schoenborn Joo Conditional UGT1A1 levels – 2009; 206: 228-233Abstract (37) harbors uridine diphosphate glycosyltransferase 1 (UGT1A1), an enzyme glucuronidation thereby elimination. Later, identified several but once again showing strongest associations [[6]Johnson A.D. Kavousi M. Smith A.V. Chen M.H. Dehghan Aspelund T. Lin Van Duijn C.M. Harris T.B. Uitterlinden A.G. Launer L. Hofman Rivadeneira Stricker B. Yang Q. Gudnason Witteman J.C. Genome-wide meta-analysis total levels.Hum. Mol. 18: 2700-2710Crossref (169) Scholar,[7]Sanna S. Busonero Maschio McArdle P.F. Usala Dei Lai Mulas Piras M.G. Perseu Masala Marongiu Crisponi Naitza Galanello R. Abecasis G.R. Shuldiner A.R. Schlessinger D. Cao Uda Common variants SLCO1B3 locus associated unconjugated hyperbilirubinemia.Hum. 2711-2718Crossref (115) Genetic very rare syndromes also frequent Gilbert syndrome, mild hyperbilirubinemia. TA-repeat polymorphism promoter accountable cases. Promoters containing seven thymine adenine (TA) repeats demonstrated be less active than six wild-type resulting lower activity higher levels, especially subjects homozygous [[8]Bosma P.J. Chowdhury J.R. Bakker C. Gantla de Boer Oostra B.A. Lindhout Tytgat G.N. Jansen P.L. Oude Elferink R.P. The basis reduced expression UDP- glucuronosyltransferase Gilbert's syndrome.N. Engl. Med. 1995; 333: 1171-1175Crossref (1219) Since antioxidative cytoprotective properties [[9]Stocker Yamamoto Y. McDonagh A.F. Glazer A.N. Ames B.N. Bilirubin antioxidant possible physiological importance.Science. 1987; 235: 1043-1046Crossref (2773) this became interesting instrument perform Mendelian randomization cardiovascular (CVD). prospective study applying (also named UGT1A1*28) not only cumulative CVD events individuals homozygote 7-TA-repeats protected against [[10]Lin Lingenhel between UGT1A1*28 allele, Study.Circulation. 2006; 114: 1476-1481Crossref (236) Many followed over contrasting results, however [11Kronenberg outcomes: more hype substance?.Circ. Cardiovasc. 2010; 3: 308-310Crossref (12) Scholar, 12Stender Frikke-Schmidt Nordestgaard B.G. Grande P. Tybjaerg-Hansen Genetically elevated risk ischaemic disease: three meta-analysis.J. Intern. 2013; 273: 59-68Crossref (64) 13Choi Lee S.J. Spiller W. Jung K.J. J.Y. Kimm Back Jee S.H. Causal decreased stroke risk: two-sample mendelian study.Arterioscler. 2020; 40: 437-445Crossref (9) 14Jeon ischemic Korea: study.Epidemiol. Health. 2019; 41e2019034Crossref (3) easy compare they different approaches. requires special genotyping procedure sophisticated simple single nucleotide (SNP). Other, recent used SNPs [[13]Choi Scholar,[14]Jeon were do focus candidate genes. Some performed extended adjustment their data, necessarily required, if instruments selected correctly. can explained fact genotype fixed time conception influenced other variables case pleiotropic effects population stratification excluded. Pleiotropy means variant influences multiple traits [[15]Holmes M.V. Ala-Korpela G.D. cardiometabolic challenges evaluating causality.Nat. Rev. Cardiol. 2017; 14: 577-590Crossref (175) Population cause confounding when structure introduced systematic differences allele frequencies across subpopulations enforced hidden frequency differs systematically [[16]Morris T.T. Davies N.M. Hemani phenomena inflate social traits.Sci. Adv. 6eaay0328Crossref (17) In pleiotropy excluded, relatively robust confounders. An even counterproductive outcome adjusted intermediate phenotype, lies causal pathway outcome. case, no statement causality given. present issue Atherosclerosis, Hou [[17]Hou Li Si Yu Sun X. Liu Yan Wang Xue Exploring Pathway Diabetes UK Biobank Cohort Study: Observational Findings Randomization Studies. 2021Google Scholar] individual data cohort 331,002 British participants summary published findings. Using approach, provided strong (CHD). Each genetically determined increase unit logarithmically transformed significantly 8% 10% CHD, respectively. diabetes could shown results hypertension inconsistent summarized data. They furthermore blood pressure, cholesterol, C-reactive protein, alcohol white cell count played important roles (Fig. 1). observational analyses, logistic regression CVD, CHD surprise, though. Only those models, age sex, pointed expected, inverse direction. All using common factors, showed increasing bilirubin. both approaches (UK data) expected Are trustable cannot reverse causation? Yes, all assumptions met [[18]Burgess Davey Dudbridge Gill Glymour M.M. Hartwig F.P. Holmes Minelli Relton C.L. Theodoratou E. Guidelines performing investigations.Wellcome Open Res. 4: 186Crossref Scholar]:1)the variants, instruments, strongly bilirubin2)the should affect confounders CVD/CHD3)the must CVD/CHD through relationship et al. plethora check assumptions. Most them exclusion effects. That is, hint anything else bilirubin? Reassuringly, methods, account such potential exclude did lead same results. Is enough effect? light this, how direction explained? show (which replicated though) factors CRP might imply effect direct mediated factors. CVD. Adjustment distort true risk. model (adjusted age, sex UB biobank center) considered informative model. How question, further investigations. One point, explain surprising definition outcomes authors applied. main reported prevalent defined ICD codes I00-I99. includes circulatory system, include cerebrovascular veins, conditions haemorrhoids varicosis. latter conditions, pathophysiological involvement absent. heterogenous mixed bag resulted finally prevalence 50%. Although whole bouquet full phenotypes prone effects, adjusting too wrong phenotypes, mediators colliders [[19]Schisterman E.F. Cole S.R. Platt R.W. Overadjustment bias unnecessary epidemiologic studies.Epidemiology. 20: 488-495Crossref (999) strengthened, though, broad definition, well tighter definitions datasets data). state-of-the-art think of. But still, questions answered yet. seems robustly shown, will essential concentrate complete taking conventional causally probably therapeutic goal, answer question does consequences. We therefore luxury address raised. declare known competing financial interests personal relationships appeared work paper. study: findings studiesAtherosclerosisVol. 320PreviewSome mildly (CVD) diabetes. Whether remains unclear. aims examine subtypes. Full-Text