cardiovascular risk factors and ischemic heart disease: is
TRANSCRIPT
DOI: 10.1161/CIRCGENETICS.115.001255
1
Cardiovascular Risk Factors and Ischemic Heart Disease: Is the Confluence
of Risk Factors Greater than the Parts? A Genetic Approach
Running title: Elosua et al.; Risk factors confluence, a genetic approach
Roberto Elosua, MD, PhD1; Carla Lluís-Ganella, PhD1; Isaac Subirana, PhD1,2;
Aki Havulinna, PhD3; Kristi Läll, MSc4; Gavin Lucas, PhD1; Sergi Sayols-Baixeras, MSc1;
Arto Pietilä, MSc3; Maris Alver, MSc4; Antonio Cabrera de León, MD, PhD5,6;
Mariano Sentí, MD, PhD7; David Siscovick, MD, MPH8; Olle Mellander, MD, PhD9;
Krista Fischer, PhD4; Veikko Salomaa, MD, PhD3; Jaume Marrugat, MD, PhD1
1Cardiovascular Epidemiology and Genetics, IMIM; 2Epidemiology and Public Health Network (CIBERESP). Barcelona, Spain; 3National Institute for Health and Welfare. Helsinki, Finland; 4Estonian Genome Center of Tartu
University. Tartu, Estonia; 5Research Unit, Nuestra Señora de la Candelaria University Hospital. Santa Cruz de Tenerife; 6University of La Laguna. La Laguna; 7Department of Experimental and Health Sciences, Pompeu Fabra University. Barcelona, Spain; 8The New York Academy of Medicine. New York, NY; 9Skåne University Hospital
Clinical Research Center. Malmö, Sweden
Correspondence:
Roberto Elosua, MD, PhD
Cardiovascular Epidemiology and Genetics
IMIM (Hospital del Mar Research Institute)
Dr Aiguader 80
08003. Barcelona, Spain
Tel: +34 933 160 800
E-mail: [email protected]
Journal Subject Terms: Epidemiology; Primary Prevention; Risk Factors; Genetic, Association Studies
, , ; , , ; , ,, ;
Krista Fischer, PhD4; Veikko Salomaa, MD, PhD3; Jaume Marrugat, MDDD, ,, PhPhPhDDD111
1CCCardiovasculararar EEpppidedeemmimiololologogogyyy ananand dd GGGeenenetete iciccss, IIMIMIM;M;M 2Epiiideeemioioololologyy aandnn PPPubuu lililic HeHeHeaalalththh Netetetwwoworkrkrk (((CICC BEBEBERERERESPSPSP).). BaBaB rrcrcelona, Spain; 333NNNationnnalll Institutututtete for HHealttth and WWeW lfarre. Helsl iininkikiki, , Finlannnd;;; 4Estonniaaan GGGeeenome CeCeenteeer oof Taartttu
UnUnUniversity. Tartuuu, EEEstonnniaaa; 5Ressseaararch Unininit,t NNNuuuestraaa SSSeñorrra de lllaaa CCCannndeeelariaaa UUUniversssityyy HHoHospspspital. SaSaSannntaaa CCrCruz ddde TeTeeneneneriririfefefe;;; 6UnUnUniviviverrssisity y y ofof LLLa LaLaL gunanana. . LaLaLa LLLaaagunnna;a;a 7DeDeDepapaparrrtttmemementntt oof ff ExExExpepeperiririmemementnn alalal aaandndnd HHeeealtltlth h ScSccieieiennncececes,s,s, PPPommmpeeeu uu FaFaFabrararaUniver isityt . Barcelona,, SSpapap in; 8Thhe e NeN w York Academy oof f Medicinee.. NeN w York, NYY; ;; 99 Skåne University Hospital
ClClClinininicicicalalal RRReseseseaeaearcrcrch h h CeCeCenntntererer... MaMaMalmmmö,ö,ö, SSSwewewedededennn
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
2
Abstract:
Background - Cardiovascular risk factors tend to aggregate. The biological and predictive value
of this aggregation is questioned and genetics could shed light on this debate. Our aim was to
reappraise the impact of risk factor confluence on ischemic heart disease (IHD) risk by testing
whether genetic risk scores (GRSs) associated with these factors interact on an additive or
multiplicative scale, and to determine whether these interactions provide additional value for
predicting IHD risk.
Methods and Results - We selected genetic variants associated with blood pressure, body mass
index, waist circumference, triglycerides, type-2 diabetes, HDL and LDL cholesterol, and IHD to
create GRSs for each factor. We tested and meta-analyzed the impact of additive (Synergy Index
–SI–) and multiplicative ( interaction) interactions between each GRS pair in one case-control
(n=6,042) and four cohort studies (n=17,794), and evaluated the predictive value of these
interactions. We observed two multiplicative interactions: GRSLDL·GRSTriglycerides interaction=–
0.096; Standard Error=0.028) and non-pleiotropic GRSIHD·GRSLDL interaction=0.091; Standard
Error=0.028). Inclusion of these interaction terms did not improve predictive capacity.
Conclusions - The confluence of LDL cholesterol and triglycerides genetic risk load has an
additive effect on IHD risk. The interaction between LDL cholesterol and IHD genetic load is
more than multiplicative, supporting the hazardous impact on atherosclerosis progression of the
combination of inflammation and increased lipid levels. The capacity of risk factor confluence to
improve IHD risk prediction is questionable. Further studies in larger samples are warranted to
confirm and expand our results.
Key words: risk factor; genetic variation; risk assessment; genetics, association studies; clustering; interactions
) p ( interaction) p
n=6,042) and four cohort studies (n=17,794), and evaluated the predictive valueee oof f f thththeseseseee
We observed nteractions. two multiplicative interactions: GRSLDL·GRSTriglycerides interaction=––
0.0909096;6;6; SStat ndndndaaard d ErErError=0.028) and non-pleiotropiiic GGGRSIHD·GRSLDDDLLL inteeraraction=0.091; Standard
EEErrooor=0.028). InInInclcllususioioion ofofof tttheheesesese inntnteereraccctiiion ttterrrms diiid nnnototot impmpmprovvve prprededediicicttivee cccaapapacaccititityyyy.
CoCoConncnclusions - Thheee confnfnflluencecece oof LDDDL L chhholeestterol aannd trtrtrigigiglyyycceridddesss geneeetiiic rririsksksk loaddd hhhasss aaan
addiitititiveveve efefeffffect oon n IHIHIHDDD iirisksk. Thhhee iininteracttioioionn bebeb twweeeenn LLLDLDLDL cchohoh llelestererrolol ananddd IHIHIHD D gggeeneneetic llloaoadd d isii
more than multiplicacacatititiveveve, susuuppppppororortitiingngng ttthehehe hhhazazazararardododoususus iiimpmpmpaccct t ononon aaathththerererosososclclclerererosososisisis ppprogression of the
co bbmbiiin ttatiiion ffof iii ffnflllamm ttatiiion anddd iiincreaseddd lilili iipiddd lllevellls. ThThThe capa iicittty offf iiri kksk fffactttor conflflfluence tttor
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
3
Introduction
The Framingham Heart Study introduced the term “cardiovascular risk factor”1 to define traits
that are associated with cardiovascular disease and have a capacity to predict future events2.
Some of these risk factors are interrelated and tend to aggregate. A paradigm of this aggregation
is metabolic syndrome3, which is associated with an increase in cardiovascular events4,5.
However, there is an open debate about whether this confluence of cardiovascular risk factors
provides clinical or mechanistic information beyond the mere addition of its individual
components6-8. In other words, is the combination of risk factors more valuable than the sum of
its parts?
An ideal way to reliably assess the impact of these risk factors on cardiovascular risk,
individually and in combination, would be to perform a prospective cohort study of individuals
with different, stable, long-term levels of exposure to these risk factors and with different
combinations of each. Alternatively, this approach could be circumvented by genetic analysis, in
which variants associated with cardiovascular risk factors are used as a proxy for the risk factors
themselves. Specifically, each risk factor could be represented by a genetic risk score (GRS)
composed of multiple variants that are known to be robustly associated with that risk factor9,10.
While this approach has the disadvantage of capturing a limited fraction of the total variance of
the risk factor itself, it does have some important advantages. First, a GRS represents constant
lifetime exposure within individuals and variable exposure between individuals, with random
combinations of alleles according to Mendel’s Second Law11. Second, it is an efficient and
economically feasible approach to this clinically important question.
In this study, we used the genetically determined variability of classical risk factors to
reappraise the value of risk factor confluence in assessing ischemic heart disease (IHD) risk. Our
An ideal way to reliably assess the impact of these risk factors on cardiovvavascscculululararar rrrisisisk,k,k,
ndividuduala lyy ana d iin n combination, would be to perfoformrm a prospective cohorort study of individuals d
wwwithhh differentntnt, stststababablee,,, lololongngng-t-t-term m m lelelevevevelslss ooof ff exxxpopoposureee ttto thhheseseseee riririsk fffacacactototorsrsr aaandndn wwititith h h dididiffffffererenenenttt
cocoommbmbinations of f eaaach. AAAlternnnatattivively,y,, tthhis apppproooachhh cccouuuldldld bbbe ciircumumumvev ntededed byy y ggegenetiicc aananaalyyysis, iiin
which variants assoccciaiaiateted withthth cccardiovasa cular riskkk factors ararareee usu ed ass a a a prprp oxy y for the risk factors
hhememseselvlveses.. SpSppececifificicalallylyy,,, eaeachch rrisisk k fafactctoror cocoululd d bebe rereprprp esesenenteted d bybyy aa gggeneneteticic rrisisk k scscorore e (G(G( RSRS) ) )
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
4
specific aims were i) to analyze whether GRSs associated with the individual cardiovascular risk
factors interact and present more than an additive or multiplicative association with IHD, and ii)
to determine whether these interactions provide additional value for predicting the risk of future
IHD events.
Methods
Design
A meta-analysis of five studies, one case-control and four prospective cohorts, was carried out.
The studies included the Myocardial Infarction Genetics Consortium (MIGen)12 and the
Framingham Heart Study (FHS), FINRISK 1997, FINRISK 2002, and Estonian Biobank
(EGCUT)13 cohorts. A total of 23,836 participants were included in the meta-analysis, 6,042
from the case-control study and 17,794 from the four cohorts.
MIGen, an international case-control study, included 2,967 cases of early-onset
and ,075 age- and sex-matched
controls (12). The FHS sample consisted of 3,557 individuals from the FHS offspring cohort
attending exam 5. Genome-wide genotype and associated phenotype data from MIGen and FHS
were obtained via the database of Genotypes and Phenotypes (dbGaP;
http://dbgap.ncbi.nlm.nih.gov; project number #5195). The FINRISK cohorts are comprised of
representative, cross-sectional population survey respondents. Surveys have been carried out
every 5 years since 1972 to assess the risk factors of chronic diseases and health behaviors in the
working age population; 5,562 individuals were included from the FINRISK 1997 cohort and
2,314 from the FINRISK 2002 cohort. Finally, the EGCUT cohort of 50,750 participants
recruited between 2002 and 2011 includes adults (aged 18-103 years) from all counties of
Estonia, approximately 5% of the Estonian average-adult population13. A subset of 6,361
Framingham Heart Study (FHS), FINRISK 1997, FINRISK 2002, and Estonian BiBiBiobobobananank k k
EGCUT)13 cohorts. A total of 23,836 participants were included in the meta-analysisi , 6,042
frrromomom the casseee-control studyyy and 17,,794 from the fououour cohorts.
MIGen, ananan inteeernnnationnnaaal case-cononontrol stttudyyy, iiincludududededed 222,9667 ccasesss ooof eeearrrly-onnnseeet
annddd ,070707555 aga e- anananddd seexx-mamatctct hhheddd
controls (12). The FFFHSHSHS sasasampmpmplelele ccconononsisisistststededed ooofff 33,3,555555777 ininindididivvviduduualalalsss frfrfromomom ttthehehe FFFHSHSHS oooffffspring cohort dd
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
5
individuals was included in the study selected for this meta-analysis.
SNP Selection
We mined published data from a series of large meta-analyses of Genome Wide Association
studies for each of the selected phenotypes. From these studies we identified SNPs that were
associated (p<5×10-8) with the trait of interest, and grouped these into 8 categories broadly
definable as distinct cardiovascular risk factors or coronary endpoints (Supplementary Table 1):
low-density lipoprotein (LDL) cholesterol14, high-density lipoprotein (HDL) cholesterol14,
triglycerides (TG)14, blood pressure (BP)15-16, type 2 diabetes (T2D)17, body mass index (BMI)18,
waist circumference19 and ischemic heart disease (IHD)20. We additionally included genetic
variants associated with schizophrenia21 as a negative control.
Genotyping
Four different arrays and two reference panels were used for genotyping and imputing. The
MIGen study used the Affymetrix 6.0 GeneChip and imputing was performed with MACH 1.0
using the HapMap CEU phased chromosomes as reference. The FHS used the Affymetrix 500K
and 50K chips, imputing was performed using HapMap CEU as reference. FINRISK used the
Illumina HumanCoreExome chip, imputation was performed using IMPUTE v222 and the 1000
Genomes Project sequencing data as a reference panel23. EGCUT used the Illumina
OmniExpress BeadChip, imputation was implemented in IMPUTE v2 using the 1000 Genomes
Project as a reference. In all the cohorts, directly genotyped SNPs were coded as 0, 1 or 2, while
the dosage was used for the imputed SNPs with values ranging between 0 and 2. SNPs with an
imputation quality <0.4 were excluded.
Construction of the genetic risk scores (GRS)
We constructed a weighted GRS for each cardiovascular risk factor of interest and IHD
variants associated with schizophrenia21 as a negative control.
Genotytyypipip ngng
FoFoFouurur differentntnt aaarrrrrrayaa ss ananand d d twtwtwo rererefefeferererencncnceee papp nenenelslsls werrre usededed fffororor ggennnoootypypypininng g g anaa d imimmpupuputititingnn . . . TTThehehe
MMMIGGeGen study yy usededed thee AAAffymmmetetetrix 6...0 0 GeGeGeneChChChip aaanddd impmpmputiiinggg waaass perfooormememed d wiw th MAMAM CHCHCH 1..0
using the HapMap CCCEUEUEU pphaseseed d d chromoosomes as rrreferencecee.. ThThT e FHS S usususedee the Affymetrix 500K
anand d 5050K K chchipipps,s,, iimpmppututining g g wawass pepep rfrforormemed d ususining g g HaHapMpMp apapp CCEUEU aass rerefefererencncee.. FIFINRNRISISK K ususeded ththee
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
6
independently by adding the number of risk alleles weighted by their effect sizes on the
phenotype of interest (Supplementary Table 1). One SNP could be included in more than one
GRS when associated with more than one risk factor, although with different weight.
We also constructed a weighted GRS for each cardiovascular risk factor, excluding those
SNPs that were related to any trait other than that of interest (non-pleiotropic GRSs). From the
list of variants associated with each trait, we excluded those that were associated with any other
trait with a p-value less than 0.10 in the Framingham cohort.
Ischemic heart disease outcomes
In the MIGen case-control study, only early-onset myocardial infarction cases were included. In
the cohort studies, two IHD outcomes were defined: hard IHD, including fatal and non-fatal
myocardial infarction and coronary death, and all IHD, additionally including angina and
revascularization. The follow-up methodology in the prospective cohorts is explained in detail in
the supplementary material. In summary, a follow-up or linkage with national databases was
implemented using predefined ICD9 and ICD10 codes. In each cohort, cases were categorized by
an event committee.
Statistical methods
The association between each GRS and IHD was tested by a logistic regression model in the
case-control study and by Cox proportional hazards models in the cohort studies. Furthermore,
we analyzed all potential pairwise interactions between the GRSs of interest and IHD. In the
analysis of these interactions –and from a methodological point of view– we considered their
departure from additivity and multiplicativity24: i) to test for multiplicative interactions we
added, one by one, all pairwise products of GRSs to the logistic or Cox regression models; and
ii) to analyze departure from additivity several metrics have been recommended, relative excess
he cohort studies, two IHD outcomes were defined: hard IHD, including fatal annndd d nonn n-n-n-fafafatatatalll
myocarardidial infn arctction and coronary death, and all IHIHD, additionally includuding angina and
eeevaaascularizatatatiooonn.n. Thhee e fofofollllllowowo -u-uup p p mememethhhodododologogogy yy in tttheee pprororospspspececectit vevee cohohohororrtststs iiis exxplplplaiaiainenened d ininin dddetetetaiaiailll ini
hhhe e susus pplementarrryyy maateteterrial. InInIn summamm rrry,,, a fofoolllow-u-u-up ororor llliiinkakakage wwwiitith natttiooonaaal dddatabbasasseses wwwas
mplemented using prprpredede efinededd IIICDCDC 9 and ICD10 cococoded s. In n eaeaeach cohort,t,t, cccasaa es were categorized by
anan eeveventnt ccomommimitttteeee..
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
7
risk due to interaction (RERI), attributable proportion (AP), and synergy index (SI)24-25. We
selected the SI metric because it has been proposed as the most robust when the model includes
covariates to control for confounding 26:
SI = [HR/ORA+B+ - 1] / [(HR/ORA+B- - 1) + (HR/ORA-B+ -1)].
Where:
- HR/ORA+B+ = Hazard ratio/Odds ratio of those exposed to factor A and B compared to
those non-exposed to factor A and B.
- HR/ORA+B- = Hazard ratio/Odds ratio of those exposed to factor A but not to factor B
compared to those non-exposed to factor A and B.
- HR/ORA-B+ = Hazard ratio/Odds ratio of those exposed to factor B but not to factor A
compared to those non- exposed to factor A and B.
This index measures the extent to which the hazard or odds ratio for both exposures together
exceeds 1, and whether this is greater than the sum of the extent to which each risk ratio,
considered separately, exceeds 1. A SI > 1 would indicate the presence of an additive interaction.
Bootstrapping was used to calculate 95% confidence intervals (95%CI) of the estimate25.
All the analyses were adjusted for age, sex, and principal genetic components to account
for population stratification and family relatedness27. We used a Bonferroni-adjusted p-value to
account for independent multiple testing. Due to the correlation between the 36 pairs of tested
interactions (each GRS of interest was included in 8 different pairwise interaction terms, we
estimated the number of effective independent tests according to the matrix of variance-
covariance28; the resulting value was 35.88. Therefore, the statistical threshold was set at
0.05/35.88=0.0014. A meta-analysis of the results observed in the different studies was
undertaken using an inverse-variance weighting under a random-effects model (DerSimonian-
- HR/ORA-R B+ = Hazard ratio/Odds ratio of those exposed to factor B but nooott tototo fffacacactototor r r A A A
cocompmppara eddd tto those non- exposed to factor AA and B.
ThThThisiss index mememeasasasurururess ttthehehe eeextxtxtenenntt t tototo wwwhihihichchch thehee hahahazarddd oor odododdsdsds rrrataa iooo fffororor bbbototothh h eexe popoosususurereresss tot gegegethththererer
exexxceeeede s 1, and wwwhhhetheeer this iiis gggreatererer thahahan thhhe summm ooof ttthehehe exxxteeent ttto wwhichhh eaccch hh risk rrratatatioio,
considered separatelelly,y,y, eexceedsdsds 111. AA SI > 1 would d d inii dicate e thththee prprp esenncecee ooof an additive interaction
BoBoBootototstststrararappppppininingg g wawawass s usususededed tttoo o cacacalclclculululatatatee e 959595%% % cococonfnfnfidididenenencecece iiintntntererervavavalslsls (((959595%C%C%CI)I)I) oooff f thththeee esesestititimamamatetete25...
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
8
Laird method)29. Heterogeneity between studies included in this meta-analysis was also analyzed
by estimating the I2 and its p-value. To assess whether an individual study had strong effects and
influenced the pooled results, a sensitivity analysis was performed by excluding one study at a
time and calculating the multiplicative and additive interaction metrics for the remaining studies.
The improvement in the predictive capacity of the statistically significant interaction
terms was evaluated by assessing improvements in discrimination and reclassification in the
cohort studies:
a) The improvement in the discriminative capacity of the model was evaluated using the
change in the c-statistic30. We first evaluated the discriminative capacity of a multivariate
model including age, sex, and all the individual GRSs of interest; additionally, we
evaluated the discriminative capacity of this multivariable model, further including the
significant interaction terms individually in different models.
b) The reclassification capacity of the interactions of interest was evaluated by calculating
the continuous net reclassification improvement index (c-NRI) and the integrated
discrimination improvement index (IDI)31-32.
These analyses were also performed in the individual studies and meta-analyzed using an
inverse-variance weighting under a random-effects model.
All statistical analyses were carried out using packaged or custom functions written in R-
3.02 (R Foundation for Statistical Computing, Vienna)33.
Ethics Statement
All participants gave written informed consent to be included in these studies. The study was
approved by the local Clinical Research Ethics Committees.
model including age, sex, and all the individual GRSs of interest; additionnanalllllly,y,, wwweee
evevaluauated d tht e discriminative capacity of thihiss multivariable model,l,, a further including the
signifffiicicananantt t innteteterararactctctioioion tetetermrmrmsss indndndivivividdduauauallll y innn dddiffefeerererentntnt mmmododdelels.s.s.
b)bb The reclaaassssificcatttion cccapappacity y y ofoff thhhe innnteeeractttiooonsss ofofof innnteeeresttt wwwas evvvaaaluaaateeed by cacaalclculullata innng
the continuooususus nen t reclclclasasassificationo impprovevevement indndndexexe (c(c-NRI)I)) ananand the integrated
didisscrcrimimininatatioion n imimprprp ovovememenent t inindedexx (I(I( DIDI)))313131 33-3222..
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
9
Results
The characteristics of the individuals included in the five studies, and the number of incident
coronary events (938 hard events and 1,453 events in total) and median follow-up in the four
cohorts, are shown in Table 1.
SNP selection and sample description
From the literature sources described above, 484 independent SNPs were reported to be robustly
associated with cardiovascular risk factors or coronary endpoints14-20. The number of SNPs
included in the GRSs ranged from 23 for Type 2 diabetes (T2D) to 81 for BMI (Supplementary
Table 1). There was a slight overlap between the different GRSs in terms of number of shared
SNPs or loci but the Spearman correlation coefficient between GRSs was weak (correlation
coefficient <0.100) with the exception of the associations between GRSs for TG and HDL -
0.391), and
(Supplementary Table 2). When the non-pleiotropic GRSs were considered only the correlation
-0.142), and remained
significant. A strong and consistent association across studies between the GRS and their
corresponding risk factors was observed, remaining strong and consistent for lipids and body
mass index when the non-pleiotropic GRS were analyzed (Supplementary Table 3).
Association between genetic risk scores and ischemic heart disease.
We observed significant associations between the GRS for IHD and hard coronary events in all
the studies and in the meta-analysis (p-value=9.4x10-47) (Figure 1 and Supplementary Table 4;
and Supplementary Figure 1 and Supplementary Table 5 for all IHD events). The TG, HDL,
LDL, BMI, and waist GRSs were also associated with coronary events in the meta-analyses of
hard IHD events, although these associations were mainly driven by the MIGen study (Figure 1
SNPs or loci but the Spearman correlation coefficient between GRSs was weak (((cocoorrrr elelelatatatioioion n n
coefficiciennt <0.10100) with the exception of the assosocic ations between GRRSsS for TG and r HDL -
0.0.0.39991), ananand d
SSSupupupplp ementary Tableee 222). WhWhWheenen the nnnoon---pleiiiotttropiic GRSRSRSs wewewere ccconnsiderereed onononlyll theee ccocorrrreellatiooonnn
-00.142),) and remained
iigngng ifificicanant.t. A A ststrorongngg anand d coconsnsisistetentnt asassosociciatatioion n acacrorossss sstutudidieses bebetwtweeeen n ththe e GRGRSS anand d ththeieir r
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
10
and Supplementary Table 4). The blood pressure, diabetes, and schizophrenia GRSs were not
associated with coronary events in this meta-analysis (Figure 1 and Supplementary Table 4).
When the non-pleiotropic GRSs were considered, only the association between the GRS for IHD
and coronary events remained significant (Supplementary Table 6).
Assessment of interactions between genetic risk scores and impact on ischemic heart
disease risk.
We tested all pairwise interactions between the GRSs of interest and IHD in the different studies.
In the meta-analyses we found two statistically significant multiplicative interactions
(Supplementary Tables 7 and 8). A negative multiplicative interaction between the LDL and TG
GRSs on all IHD events (Table 2 and Supplementary Table 8). When hard IHD events were
considered, the magnitude of the association of the interaction term decreased, from -0.096 to -
0.047) (Table 3), but this decrease was driven by the MIGen study; when that study was
excluded the effect of the interaction term on hard IHD remained similar and statistically
significant -0.116; p-value=1.3x10-4) (Supplementary Table 9). A positive multiplicative
interaction between the non-pleiotropic LDL and IHD GRSs on all IHD and hard IHD was also
observed (Table 2), and was robust and consistent in the sensitivity analysis (Supplementary
Table 9).
We also analyzed the presence of additive interactions. In the meta-analysis, we did not
find any statistically significant additive interaction term (Supplementary Tables 10 and 11).
We estimated 80% statistical power to detect a multiplicative interaction regression
coefficient higher or lower than ± 0.077, considering the observed standard error (0.020) and a p-
value=0.0014. We also estimated 80% power to detect a synergy index higher than 1.28 or lower
than 0.72, considering the lower observed standard error (0.07), and a synergy index higher than
GRSs on all IHD events (Table 2 and Supplementary Table 8). When hard IHD eevevenee tstss wwwererere e e
considere ede , thhe mamagnitude of the association of the e ini teraction term decrereasa ed, from -0.096 to -
0.0.0.04447) (Tableee 333))),, bubb t t thththisisis dddecee reeeasasaseee wawawas s s drdrdrivvvenenen bby tthheee MIMIIGeGeGen n n sts ududdy;y;; wwwheheennn thtt att stststudududy y y wawaasss
exexxcllluduu ed the effececect of thhhe intteterraraction n n teeermrmm onnn hhhard IHDHDHD rereremaaiinedd siiimilarrr aaandd d stststatistiiicacaallyyy
ignificant -0.116166;;; p-p-p value=e=e=1.11 3x10-4)) (S( uppplp emmmene tary TTTababablee 9)). A popoposiss tive multiplicative
nnteteraractctioion n bebetwtweeeen n ththe e nonon-n plplp eieiototroropipip c c LDLDL L anand d IHIHDD GRGRSsSs onon aallll IHIHDD anand d hahardrd IIHDHD wawas s alalsoso
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
11
2.57 or lower than -0.57, considering the higher observed standard error (0.39), always with a p-
value=0.0014.
Assessment of the predictive capacity of the scores
We evaluated improvement in the discrimination of coronary events in the different cohort
studies. First we used a model that included age, sex, the GRSs for all the cardiovascular risk
factors evaluated, and the first two principal genetic components. Second, we added to the
model, the interaction terms that were associated with IHD (GRSLDL·GRSTG, and non-pleiotropic
GRSIHD·GRSLDL). Including these interaction terms did not improve the discriminative or
reclassification capacity of coronary events in the meta-analysis (Table 3).
Discussion
In the present study, we evaluated the potential interaction effects between cardiovascular risk
factors on ischemic heart disease risk using a genetic approach. We tested the departure from an
additive or multiplicative effect of the different two-pair combinations of GRSs related to these
risk factors and their association with coronary events. We report two significant multiplicative
interactions (GRSLDL·GRSTG and non-pleiotropic GRSIHD·GRSLDL) modulating coronary risk.
The inclusion of these interaction terms in the multivariate model did not improve the predictive
capacity of the model based on the individual effects of the GRSs of interest.
We first evaluated the association of each individual GRS with its corresponding risk
factors and these associations were strong and consistent across studies. We also evaluated the
effects of each individual GRS on IHD risk in each study and meta-analyzed the results. The
GRS for IHD was associated with coronary events in all the studies and also in the meta-
analysis. The GRSs for the different risk factors were also associated with hard coronary events
in the meta-analysis, with the exception of the GRSs for blood pressure, diabetes, and
Discussion
nnn tthehehe presentntnt study, we evaluated the pop tential inttterrraction effectsss bebb tween cardiovascular risk
fafafactttoro s on ischemmmiiic heeearrrt disseaeaase risk uuusiiing aaa gggeneeeticc apppprprproaoaoachh. WeWeWe testededed thhehe dddepartrtrtuurre frrromo annn
addiitititiveveve ooorr multltltiiiplililicacatititiveve effeccttt ofofof the didiiffffffeerenent twtwoo-ppaiaiirr ccocombmbmbiinatttioioionsns oofff GGRGRSss tttheheh seserererellalated dd ttoto
isk factors and theiiirrr asasassososociciciatatatioioionnn wiwiwiththth cococorororonananaryryry eeeveveventntntss.s. WeWee rererepopoportrtrt twtwtwooo sisisigngngnifififiiicacacant multiplicative
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
12
schizophrenia (which was included as a negative control). These results validate the GRSs; the
lack of association of IHD events with blood pressure and diabetes could be related to the lack of
causal relationship11, low statistical power in the prospective studies, or other factors.
The debate about whether the aggregation of cardiovascular risk factors provides
additional information on vascular health beyond that of each individual components is still
open. The paradigm for this discussion is metabolic syndrome. Our choice of a genetic approach
to assess whether different risk factors interact to modulate the risk of IHD was based on the
premise that a genetic score for a given risk factor captures some of its population variability;
however, the extent to which this is true varies markedly between risk factors. The amount of
variance in the traits of interest that is accounted for by genetic scores varies from ~25% to 30%
for LDL cholesterol14 down to no more than 3% for blood pressure15. However, the loss of
information that this represents, with respect to measuring the phenotype itself, is
counterbalanced by the fact that genetic risk is a constant exposure throughout an individual’s
lifespan. Some studies have suggested that selecting a list of SNPs nominally associated with a
trait increases the explained variability of that trait34. In this study, we selected only those SNPs
consistently replicated in GWAS to be associated with the phenotypes of interest. The allelic
scores that include thousands of genetic variants tend to lack specificity, and therefore should be
used with caution and perhaps only to analyze proxy biological intermediates, not to analyze the
association with other related clinical phenotypes34, as in the present study. Moreover, the list of
nominally associated SNPs could vary across studies. For all these reasons, we preferred to select
those variants with a statistically significant association, considering the GWAS threshold for
our analyses.
In the analysis of these interactions we considered their departure from additivity and
variance in the traits of interest that is accounted for by genetic scores varies frommm ~2~2~25%5%5% ttto o o 303030%%
for LDDL L chc ollese teerrolo 14 down to no more than 3% foror blood pressure15. Hoowew ver, the loss of
nnnfooormation thththatatat ttthihh ss rerereprprpresesesenee tstss, , , wiwiwiththth rrresesespepp ctctct tttooo meeeaaasurinining g g thththe e phphhenenenotototypypypeee itii selflff, , isisis
cocooununnterbalanced d bby thhee ffact tththataa geneneeticcc rrrisk isss a cooonnnstaaantntnt exxpxposurrre througugughohohoututut an innndidiivividuduual’sss
ifespan. Some studdieieies s s have ssuguguggegeg sted tthat selecttinining gg a list ooof f f SNSNPs nomomminininalff lyy associated with a
rraiait t inincrcreaeasesess ththe e exexplplp aiainened d vavaririababililitity y y ofof tthahat t trtraiaitt343434.. InIn tthihis s ststududy,y,y, wewe sselelececteted d ononlylyy tthohosese SNSNPsPs
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
13
multiplicativity24 and identified two multiplicative interaction terms, one showing a less than
multiplicative effect (GRSLDL·GRSTG) and other a more than multiplicative effect (non-
pleiotropic GRSLDL·GRSIHD). The LDL and TG GRSs were slightly correlated. This association
could be related to common molecular mechanisms or to the use of the Friedewald equation to
estimate LDL in most epidemiological studies. Although this collinearity could decrease the
statistical power of our analyses, we report a statistically significant multiplicative interaction
between the genetic load for LDL cholesterol and TG. This interaction term had a negative
value, indicating that the joint effect of these two factors is less than multiplicative in the risk
ratio scale. Moreover, as the additive interaction between these two factors was not statistically
significant, we can assume an additive effect of these two factors on IHD risk in the risk ratio
scale. This type of additive but not multiplicative effect of two risk factors has also been reported
in other diseases, e.g., to describe the joint effects of smoking and asbestos on lung cancer35. The
explanation of this additive effect could be related to basic lipid profile concepts36. The lipid
profile includes measurement of the total amount of the two most important lipids in the plasma
compartment: cholesterol and TG. These lipids are not soluble in plasma, and are carried in
association with proteins, the so-called lipoproteins: HDL, LDL and TG-rich lipoproteins. The
TG-rich lipoproteins also transport remnant cholesterol. Triglycerides can be degraded by most
cells, but cholesterol cannot; therefore, the cholesterol content of TG-rich lipoproteins, rather
than increased TG levels per se, is the more likely contributor to atherosclerosis and
cardiovascular disease36. The negative multiplicative interaction indicates an additive effect
between TG and LDL cholesterol on IHD risk, and supports the suggestion that TG-rich particles
act as an additional source of cholesterol in the arterial wall.
We also report a more than multiplicative effect between the non-pleiotropic genetic load
ignificant, we can assume an additive effect of these two factors on IHD risk in tthehehe risisisk k k rararatititio o
cale. ThThisi typypype ofof additive but not multiplicative eeffffect of two risk factoorsrs has also been reported
nnn ooottther diseaaasesees,s,s, eee.gg...,,, tototo dededescs ririribebebe tttheheh jjjoioiointnn eeeffffffecee ts offf smmmokokokinining gg annnddd asasasbebebestststosoo on n lululungngng cccannncececerrr355. . TTTheh
exexxplpllaanation of thhhisss adddditttive eefefffefect couououlddd bbbe reeelaaatedd tooo bbasasasiciic lllipppid ppprooofile cococonccepepeptst 36. ThThThe llipppid
profile includes meaaasususurerementtt ooof ff the totat l amountt ooof the twwwo o o momost impmpororortatat nt lippids in the plasma
cocompmpparartmtmenent:t: cchoholeleststererolol aandnd TGTG.. ThThesesee lilipipip dsds aarere nnotot ssololubublele iin n plplp asasmama, , , anand d arare e cacarrrrieied d inin
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
14
for LDL and IHD. The IHD genetic load has been related to lipid, inflammatory and immune
pathways that could potentiate the progression of atherosclerosis37. The non-pleiotropic GRS for
IHD excluded SNPs associated with lipids and mainly reflects inmuno-inflammatory
mechanisms. Therefore, this interaction could be explained by the independent interrelationships
between lipids and inmuno-inflammation that could trigger the deleterious consequences of these
two factors through different mechanisms38,39.
We also analyzed the improvement in predictive capacity when the interaction terms were
included in the model. However, we did not observe any improvement in the discrimination or
reclassification. Recent meta-analyses focused on metabolic syndrome have shown that the
population with this syndrome has a two-fold higher risk of cardiovascular disease than the rest
of the population4-5 but the added value of this clinical constellation of risk factors is questioned6-
8. We identified 1 cross-sectional study40 and 6 cohort studies41-46 that assessed the unadjusted
and adjusted association between metabolic syndrome and cardiovascular risk. When the models
were adjusted for all or some of the classical cardiovascular risk factors, 3 of these studies
showed an association between metabolic syndrome and cardiovascular events43,45-46. However,
Girman et al only adjusted for the estimated coronary risk obtained with the Framingham
45, and McNeill et al did not adjust for HDL cholesterol
and BP46. In contrast, our analyses did not show any interaction between the GRSs related to the
risk factors that define metabolic syndrome. Our results are in line with the two remaining
studies, which specifically analyzed whether metabolic syndrome improves the predictive
capacity of its individual components. Neither study reported significant improvement in
discrimination capacity44,46; this shared finding calls into question the capacity of the metabolic
syndrome diagnosis to improve a cardiovascular risk calculation based on the individual classical
population with this syndrome has a two-fold higher risk of cardiovascular diseasassee thththananan ttthehehe rrrest
of the pppopoppulatationn4-5 but the added value of this clininicac l constellation of risisk factors is questioned6
... WWWe identifffieiied d d 111 crosososs-s-s-sesesectctctionananal ll stststudududyyy4040 andndnd 666 cohhhooort ststtudududieieiesss41-44666 ththhatata asasassesesessededd ttthehehe uuunanaadjdjdjusususteteteddd ddd
anannd d adaa justed assococociationonon betweweweene metete abbbooolic sssyyyndromomome anananddd ccacarrrdiooovaaasculaaar risssk.k. Wheeen n ththee mmmodddellls
were adjusted for all l l ororor some e ofofof the classs ical cardididiovovasculalaar r r rirr sks factoorsrsrs,,, 3 of these studies
hhowoweded aan n asassosociciatatioion n bebetwtweeeen n memetatabobolilic c sysyyndndroromeme aandnd ccarardidiovovasascuculalar r evevenentsts43,45-46.. HoHowewevever,r,,
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
15
cardiovascular risk factors.
Limitations of the study
Four main limitations should be considered: i) The variability of the cardiovascular traits
explained by the genetic scores considered in this analysis is not very high, in general, but
represents lifetime exposure. Moreover, some interacting genetic variants could have been
overlooked by GWAS and therefore not included in our GRSs; ii) The small number of events
observed in the cohort studies limited the statistical power to explore the interactions of interest.
We have also to consider that when the magnitude of the association between the two individual
components of the interaction and the outcome of interest is small the power to differentiate
between additive and multiplicative effects is reduced; iii) IHD clinical endpoints are the result
of a complex phenomenon, which includes endothelial dysfunction, plaque formation and
growth, plaque stability, and thrombosis. Interaction could happen in the context of one of these
pathways and be diluted in the observation of clinical end-points; and iv). Although the approach
we used could be considered as Mendelian randomization11, we must be cautious about
interpreting the causality and synergistic effect of the confluence of risk factors. First, the genetic
instrumental variable is a genetic score composed by multiple risk alleles10. In some cases, the
biological pathway linking each risk allele to the intermediate trait of interest is unknown, and
therefore the assumption that the only causal pathway from the genetic variant to IHD involves
the trait of interest is questionable. Moreover, there may be association(s) between the genetic
variants and unmeasured/unknown confounders; for example, the genetic load of obesity could
be related to food choices that could also be directly related to coronary risk. We also must
consider the presence of pleiotropic effects that are reflected in the correlation between the GRSs
analyzed and that violate one of the assumptions of Mendelian randomization studies.
between additive and multiplicative effects is reduced; iii) IHD clinical endpointtss s ararare ththhe e e rereresusus ltll
of a comomplplp exx ppheenon menon, which includes endothhelelial dysfunction, plaququq e formation and
grgrgrowwwth, plaqueueu ssstatatabililiitytyty,,, ananand dd thhhrororombmbmbosossisisis.. Innnteteterarar ctiooonnn cooulululd dd hahahapppenenen iiin n n thhhee e cocc nttexexext tt ofofof oonenee ooof f f thththesesese e
papaathhhwaw ys and beee dddiluteeeddd in thehehe obserrrvavav tttiooon ooof cccliniiicaaal eeendndnd-poiiints;;; aaandn ivv). Alllththhouo ghh thhhe apppproooaccch
we used could be conononsisis ded red asasas Mendelian randoomimim zationn11,,, wewe mustt bebebe cautious about
nnteterprppreretitinng g g ththe e cacaususalalitity y y anand d sysyynenergrggisistitic c efeffefectct oof f ththe e coconfnfluluenencece oof f ririsksk ffacactotorsrs.. FiFirsrst,t,, tthehe gggeneneteticic
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
16
Finally, we would note that 339 of the FINRISK participants were also included in the
MIGen sample; however, this is a small proportion (<1.5%) of the whole sample, the sensitivity
analyses carried out are consistent, and we could consider the effect of this duplication to be
minimal.
Conclusions
The genetic risk loads for LDL cholesterol and TG interact, suggesting that the effect of these
two risk factors on IHD risk is additive rather than multiplicative. Moreover, the non-pleiotropic
GRSs for LDL and IHD also interact on IHD risk and have a more than multiplicative effect.
This interaction supports the hazardous impact on atherosclerosis progression of the combination
of inflammation and increased lipid levels. Our results question the added value of the
confluence of risk factors in improving the estimation of cardiovascular risk beyond the
predictive capacity provided by individual risk factors. However, further studies in larger
samples are warranted to confirm and expand our results, due to the limited statistical power of
the present analysis.
Acknowledgments: Access to the Myocardial Infarction Genetics Consortium (MIGen) and
Framingham data was provided through the Database of Genotypes and Phenotypes (dbGaP;
http://dbgap.ncbi.nlm.nih.gov; project number #5195). The MIGen Consortium was funded by
grant R01 HL087676 (NIH, USA). The Framingham Heart Study is conducted and supported by
the National Heart, Lung, and Blood Institute (NHLBI) in collaboration with Boston University
(Contract No. N01-HC-25195). This manuscript was not prepared in collaboration with
investigators of the Framingham Heart Study and does not necessarily reflect the opinions or
views of the Framingham Heart Study, Boston University, or NHLBI. Funding for SHARe
genotyping was provided by NHLBI Contract N02-HL-64278. To Elaine M. Lilly, PhD,
(Writer's First Aid) for her critical reading and revision of the English text.
This interaction supports the hazardous impact on atherosclerosis progression off ttthehehe cccomomombibibinananatitition
of inflammation and increased lipid levels. Our results question the added value of f hthe
cooonfnfnfllluence oooff f risk factors in imprp oving g the estimatttiooon of cardiovaaascscscular risk beyoyy nd the
prprpredddictive capaciiityyy provovovided bbby y individdduaaal riiskkk faccctooors. HHHowowowevvver, fufufurtherrr ssstududdieiees in lllararrgegeer
ampplelelesss aararee waarrrranantteted dd ttoto cconfifiirmrmm and expxppanand dd ouour r reresusultltts,ss dddueee ttto thehehe lllimi ititit deded statititistststiciical ppowowerer oofff
he present analysis.
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
17
Sources of Funding: This work was supported by the Spanish Ministry of Economy and
Innovation through the Carlos III Health Institute [Red HERACLES RD12/0042, CIBER
Epidemiología y Salud Pública, PI09/90506, PI12/00232], European Funds for Development
(ERDF-FEDER), and by the Catalan Research and Technology Innovation Interdepartmental
Commission [SGR 1195]. GL was funded by the Juan de la Cierva Program, Ministry of
Education (JCI-2009-04684). SSB was funded by an iPFIS contract, Carlos III Health Institute
(IFI14/00007). FINRISK: The FINRISK surveys were mainly funded from budgetary funds of
Finland’s National Institute for Health and Welfare. Important additional funding has been
obtained from the Academy of Finland (grant # 139635) and from the Finnish Foundation for
Cardiovascular Research. EGCUT received financing from FP7 grants (278913, 306031,
313010), an Estonian Research Council Grant (GP1GV9353), the Center of Excellence in
Genomics (EXCEGEN), and the University of Tartu (SP1GVARENG).
Disclosures: None.
References:
1. Kannel WB, Dawber TR, Kagan A, Revotskie N, Stokes J III. Factors of risk in the development of Coronary Heart Disease—Six-Year Follow-up Experience. The Framingham Study. Ann Intern Med. 1961;55:33–50.
2. O’Donnell CJ, Elosua R. Cardiovascular risk factors. Insights from Framingham Heart Study. Rev Esp Cardiol. 2008;61:299–310.
3. Alberti KG, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donate KA, et al. Harmonizing the metabolic syndrome. Circulation. 2009;120:1640–1645.
4. Gami AS, Witt BJ, Howard DE, Erwin PJ, Gami LA, Somers VK, et al. Metabolic syndrome and risk of incident cardiovascular events and death. A systematic review and meta-analysis of longitudinal studies. J Am Coll Cardiol. 2007;49:403–414.
5. Mottillo S, Filion KB, Genest J, Joseph L, Pilote L, Poirier P, et al. The metabolic syndrome and cardiovascular risk a systematic review and meta-analysis. J Am Coll Cardiol.2010;56:1113–1132.
6. Reaven GM. The metabolic syndrome: requiescat in pace. Clin Chem. 2015;51:931–938.
( ), y ( )
Disclosures: None.
RRReffeferences:
1. KKKanana nel WBWBWB, DaDaDawbeeer TR, KKKagagagann AAA, ReReRevootttskkkie NN,N, Stototokkkes J J III. FFaaactorsss ooof riririskskk in thththeee develololopmpmpmeenent ofoff CCCorororonononarararyyy Hearrrttt DiDiDisease—ee SiSiSixxx-YeYeYeararar FFFololollololowww-upupup Expxpxpeererieeencncnceee. Thehehe FFFrararamingngnghahahammmStudy. Ann Intern MMMededed. 191919616161;5;5;55:5:5:333333–5–5500.0.
222. OOO’D’DDonononnenenellllll CCCJJJ, EEElololosususuaaa RRR. CCCararardididiovovovasasascucuculalalarrr riririsksksk fffacacactototorsrsrs. InInInsisisighghghtststs fffrororommm FrFrFramamaminininghghghamamam HHHeaeaeartrtrt SSStututudydydy. C d l
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
18
7. Kahn R, Buse J, Ferrannini E, Stern M. The metabolic syndrome: time for a critical appraisal. Joint statement from the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care. 2005;28:2289–2304.
8. Woodward M, Tunstall-Pedoe H. The metabolic syndrome is not a sensible tool for predicting the risk of coronary heart disease. Eur J Cardiovasc Prev Rehabil. 2009;16:210–214.
9. Hindorff LA, MacArthur J, Morales J, Junkins HA, Hall PN, Klemm AK, et al. A Catalog of Published Genome-Wide Association Studies. Available at: www.genome.gov/gwastudies. Accessed [15-01-2013].
10. Burgess S, Thompson SG. Use of allele scores as instrumental variables for Mendelian randomization. Int J Epidemiol. 2013;42:1134-1144.
11. Burges S, Timpson NJ, Ebrahim S, Davey Smith G. Mendelian randomization: where are we now and where are we going? Int J Epidemiol. 2015;44:379-388.
12. Kathiresan S, Voight BF, Purcell S, Musunuru K, Ardissino D, Mannucci, et al. Genome-wide association of early-onset myocardial infarction with single nucleotide polymorphisms and copy number variants. Nat Genet. 2009;41:334–341.
13. Leitsalu L, Haller T, Esko T, Tammesoo ML, Alavere H, Snieder H, et al. Cohort Profile: Estonian Biobank of the Estonian Genome Center, University of Tartu. Int J Epidemiol.2014;44:1137-1147.
14. Global Lipids Genetics Consortium, Willer CJ, Schmidt EM, Sengupta S, Peloso GM, Gustafsson S, et al. Discovery and refinement of loci associated with lipid levels. Nat Genet.2013;45:1274-1283.
15. Ehret GB, Munroe PB, Rice KM, Bochud M, Johnson AD, Chasman DI, et al. Genetic variants in novel pathways influence blood pressure and cardiovascular disease risk. Nature.2011;478:103–109.
16. Wain LV, Verwoert GC, O’Reilly PF, Shi G, Johnson T, Johnson AD, et al. Genome-wide association study identifies six new loci influencing pulse pressure and mean arterial pressure. Nat Genet. 2011;43:1005–1011.
17. Mahajan A, Go MJ, Zhang W, Below JE, Gaulton KJ, Ferreira T et al. Genome-wide trans-ancestry meta-analysis provides insight into the genetic architecture of type 2 diabetes susceptibility. Nat Genet. 2014;46:234-244.
18. Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH, Day FR, et al. Genetic studies of body mass index yield new insights for obesity biology. Nature. 2015;518:197-206.
19. Shungin D, Winkler TW, Croteau-Chonka DC, Ferreira T, Locke AE, Mägi R, et al. New genetic loci link adipose and insulin biology to body fat distribution. Nature. 2015;518:187-196.
12. Kathiresan S, Voight BF, Purcell S, Musunuru K, Ardissino D, Mannucci, et alll. . GeGeGenonomeme-wide association of early-onset myocardial infarction with single nucleotide polyyymomoorppphihihismsmsmsss and copy number variants. Nat Genet. 2009;41:334–341.
133.. LeLeLeiititsalululu LLL, HaHaHalller T, Esko T, Tammesoo ML,L AAAlllavere H, Sniededd r H,H eeet t al. Cohort Profile: EsEsEstooonian Bioobababanknknk of f thththeee EsEsEstott nininiananan GGGenennomomome CeCeCentnn er,,, UUUnivvvererersisisitytyty of f f TaTaTartrtrtu.u.. InInInt J EpEpEpidididemememioolll..2202 1114;44:1137-11414147.
144. GlGlGlobobobalalal LLLiipipididids GeGeGennnetititicscc CConononsosortrtrtiiuium,m,m, WWWililillelelerrr CJCJCJ,,, ScScSchmhmhmidididt tt EMEMEM, SeSeSengngngupupuptaaa SS, PePePelololosososo GGGM,M,M, Gustafsson S, et al. DDDisisi coc veryryy aaandndn refinement of lllococo i assoociciciataa eded with lililipipipid levels. Nat Genet.2013;45:1274-12833.3.
1515 EhEhrerett GBGB MuMunrnroeoe PPBB RRicicee KMKM BoBochchudud MM JoJohnhnsosonn ADAD ChChasasmamann DIDI etet alal GeGenenetiticc
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
19
20. Nikpay M, Goel A, Won HH, Hall LM, Willenborg C, Kanoni S, et al. A comprehensive 1,000 Genomes-based genome-wide association meta-analysis of coronary artery disease. Nat Genet. 2015;47:1121-1130.
21. Schizophrenia Working Group of the Psychiatric Genomics Consortium. Biological insights from 108 schizophrenia-associated genetic loci. Nature. 2014;511:421-427.
22. Howie BN, Donnelly P, Marchini J. A flexible and accurate genotype imputation method for the next generation of genome-wide association studies. PLoS Genet. 2009;5:e1000529.
23. 1000 Genomes Project Consortium, Abecasis GR, Auton A, Brooks LD, DePristo MA, Durbin RM, et al. An integrated map of genetic variation from 1,092 human genomes. Nature.2012;491:56-65.
24. VanderWeele TJ, Knol MJ. A Tutorial on Interaction. Epidemiol Method. 2014;3:33–72.
25. Knol MJ, Van der Tweel I, Grobbee DE, Numans ME, Geerlings MI. Estimating interaction on an additive scale between continuous determinants in a logistic regression model. Int J Epidemiol. 2007;36:1111–1118.
26. Skrondal A. Interaction as departure from additivity in case-control studies: A cautionary note. Am J Epidemiol. 2003;158:251–258.
27. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet.2006;38:904-909.
28. Cheverud JM. A simple correction for multiple comparisons in interval mapping genome scans. Heredity. 2001;87:52–58.
29. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177–188.
30. Newson R. Confidence intervals for rank statistics: Somers’ D and extensions. Stata Journal.2006;6:309–334.
31. Pencina MJ, D’Agostino Sr. RB, Steyerberg EW. Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers. Stat Med. 2011:30:11–21.
32. Chambless LE, Cummiskey CP, Cui G. Several methods to assess improvement in risk prediction models: Extension to survival analysis. Stat Med. 2011;30:22–38.
33. Team RDC. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 2010.
25. Knol MJ, Van der Tweel I, Grobbee DE, Numans ME, Geerlings MI. Estimatinnng g g innteteraractctioion on an additive scale between continuous determinants in a logistic regression mooodededell.l. InInInt t t J J J Epidemiol. 2007;36:1111–1118.
266.. SkSkSkror ndnddalalal AAA. InInInteraction as departure from adddididitivvity in case-cononntrol ssstututudies: A cautionary nononoteee. Am J EEpipip dededemimm ololl.. 202020030303;1; 585858:2:2:2515151–2–2–2585858.
27277. PrPP ice AL,, Paaatteeersoonn NJ, PPlPlenene ge RRRMMM, WWeiiinbbblatttt MMME,E,E, SSS ahahadiiick NNNAAA, Reeeiccch DDD. . Princcicipppalllcompmpmponononenenentststs aanananalylylysisisis s s cccorrrrrrecee tss fffororo ssstrtrtraatiffficicicatatatioioion n n ininin gggennnomomome-e-e wiwiwidedede assssssococociaiaiatititiononn ssstuuudididieseses.. NNNat t t GeGeGenenenettt.2006;38:904-909.
2828. . ChChevevererudud JJM.M. AA ssimimplplp e e cocorrrrecectitionon fforor mmulultitiplplp ee cocompmppararisisonons s inin iintnterervaval l mamappppppining g g gegeg nonomeme ccananss HeHereredidityty 20200101;8;87:7:5252 5–588
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
20
34. Evans DM, Brion MJ, Paternoster L, Kemp JP, McMahon G, Munafò M, et al. Mining the human phenome using allelic scores that index biological intermediates. PLoS Genet.2013;9:e1003919.
35. Hilt B, Langard S, Lund-Larsen PG, Lien JT. Previous asbestos exposure and smoking habits in the county of Telemark, Norway - A cross-sectional population study. Scand J Work Environ Health. 1986;12:561–566.
36. Nordestgaard BG, Varbo A. Triglycerides and cardiovascular disease. Lancet.2014;384:626–635.
37. CARDIoGRAMplusC4D Consortium, Deloukas P, Kanoni S, Willenborg C, Farrall M, Assimes TL, et al. Large-scale association analysis identifies new risk loci for coronary artery disease. Nat Genet. 2013;45:25-33.
38. Shapiro MD, Fazio S. From Lipids to Inflammation: New Approaches to Reducing Atherosclerotic Risk. Circ Res. 2016;118:732-749.
39. Tall AR, Yvan-Charvet L. Cholesterol, inflammation and innate immunity. Nat Rev Immunol. 2015;15:104-116.
40. Alexander CM, Landsman PB, Teutsch SM, Haffner SM. NCEP-defined metabolic syndrome, diabetes, and prevalence of coronary heart disease among NHANES III participants age 50 years and older. Diabetes. 2003;52:1210-1214.
41. Wilson PW, D'Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation. 1998;97:1837-1847.
42. Yarnell JW, Patterson CC, Bainton D, Sweetnam PM. Is metabolic syndrome a discrete entity in the general population? Evidence from the Caerphilly and Speedwell population studies.Heart. 1998;79:248-252.
43. Sattar N, Gaw A, Scherbakova O, Ford I, O’Reilly DS, Haffner SM, et al. Metabolic syndrome with and without C-reactive protein as a predictor of coronary heart disease and diabetes in the West of Scotland Coronary Prevention Study. Circulation. 2003;108:414-419.
44. Stern MP, Williams K, González-Villalpando C, Hunt KJ, Haffner SM. Does the metabolic syndrome improve identification of individuals at risk of type 2 diabetes and/or cardiovascular disease? Diabetes Care. 2004;27:2676–2681.
45. Girman CJ, Rhodes T, Mercuri M, Pyörälä K, Kjekshus J, Pedersen TR, et al. The metabolic syndrome and risk of major coronary events in the Scandinavian Simvastatin Survival Study (4S) and the Air Force/Texas Coronary Atherosclerosis Prevention Study (AFCAPS/TexCAPS). Am J Cardiol. 2004;93:136-141.
39. Tall AR, Yvan-Charvet L. Cholesterol, inflammation and innate immunity. NaNaNat t t ReReRev v v Immunol. 2015;15:104-116.
400.. AlAlAlexe anndeded rr r CMCMCM, Landsman PB, Teutsch SM, HHHaffffner SM. NCEEEP-P defififinnned metabolicyyyndddrome, diiiabababetetetesee , ananand dd prprpreveve alallenenencecece oof ff cococorooonananaryrr heaeaearrt ddisisiseaeaeasesese amomomongngng NNNHAHAHANEESSS IIIIIII I I papap rtrtticicicipipipananantststs
agaga ee 50 years anddd ooolderrr. DDiDiababetetetes. 202020030 ;5;52:1222110-12121214.
411. WiWiWilslslsononon PPPW,W,W, DDD'AAAgogogostststininino RBRBRB,,, LeLeLevy DDD, ,, BeBeBelalaangngngeerer AAAM,M,M, SSSilililbebebersrsr hahaatztzz HHH,,, KaKaKannnnn elelel WWWB.B.B. PPrererediiictctc ioioion of coronary heart diseseseasasa e usinnng g g rirr sks facctot r categogoriririese . Circcculululatata ioi n. 19989898;9;9;97:1837-1847.
4242. . YaYarnrnelell l JWJW, , , PaPatttterersoson n CCCC, , , BaBainintoton n D,D,, SSweweetetnanam m PMPM.. IsIs mmetetababololicic ssynyny drdromome e a a didiscscreretete enentitityty iinn ththee gegenenerarall popopupulalatitionon?? EvEvididenencece ffroromm ththee CaCaererphphilillyly aandnd SSpepeededwewellll ppopopululatatioionn ststududieiess
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
21
46. McNeill AM, Rosamond WD, Girman CJ, Golden SH, Schmidt MI, East HE, et al. The metabolic syndrome and 11-year risk of incident cardiovascular disease in the atherosclerosis risk in communities study. Diabetes Care. 2005;28:385–390.
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
22
Table 1: Characteristics of the Participants in the Five Studies Included in the Meta-analysis (Number and Percentage Shown for Categorical Variables, Mean and Standard Deviation for Continuous Variables).
Study MIGen* FHS* FINRISK1997 FINRISK2002 EGCUT*
Design Case-control Cohort Cohort Cohort Cohort
N Hard IHD events
6,0422967
3,557168
5,562367
2,314145
6,361258
All IHD events --- 251 447 185 570
Follow-up, years (SD) --- 12.8 (6.2) 13.8 (2.9) 9.2 (1.9) 5.4 (2.4)
N women (%) 1,422 (23.54%) 1,880 (52.9%) 2,878 (51.7%) 1,106 (47.8%) 3,615 (56.8%)
Age, in years NA* 54.61 (9.8) 47.9 (13.3) 51.7 (12.6) 48.2 (19.4)
SBP*, in mmHg NA 125.8 (18.6) 135.8 (19.7) 137.9 (20.4) 128.6 (18.1)
DBP*, in mmHg NA 74.5 (9.9) 82.3 (11.3) 80.2 (11.3) 78.7 (10.9)
BMI*, in kg/m2 NA 27.4 (5.0) 26.6 (4.5) 27.5 (5.1) 26.5 (5.2)
HTN*, n (%) NA 1,020 (28.7%) 2,477 (44.5%) 1,143 (49.4%) 1,771 (27.8)
LDL cholesterol*, in mg/dL NA 125.1 (34.5) 134.8 (35.9) 138.2 (48.7) 137.3 (40.8)†
HDL cholesterol*, in mg/dL NA 50.1 (15.1) 54.8 (13.5) 58.7 (18.9) 58.8 (15.9)†
Triglycerides, in mg/dL NA 139.4 (78.6) 130.1 (90.3) 137.2 (92.9) 139.2 (90.9)†
Current Smoking, n (%) NA 610 (17.2%) 1,355 (24.4%) 617 (26.7%) 1,904 (29.9%)
* MIGen: Myocardial Infarction Genetics Consortium; FHS: Framingham Heart Study; EGCUT: Estonian Biobank; SBP: Systolic blood pressure; DBP: Diastolic blood pressure; BMI: Body mass index; HTN: Hypertension; LDL: Low-density lipoprotein; HDL: High-density lipoprotein; NA: Not available.† Data available in a subsample of 3,782 individuals.
HD events --- 251 447 185 575757000
w-up, years (SD) --- 12.8 (6.2) 13.8 (2.9) 9.2 (1.9) 555.444 (2(2(2 44.4) ))
men (%(%))) 1,422 (23.54%) 1,880 (52.9%%)) ) 2,878 (51.7%) 1,1,1,1011 6 (47.8%) 3,615 (56.8%
in n n yeyeeaaars NA* 54.61 (9.8)) ) 47.9 (13.3))) 551.7 (12.6) 48.2 (19.4)
,, innn mmmHg NANANA 125.5..8 ((18.6)) 1355.5.888 (19.99 7)77 13337.9 (2(2(200.0.4)4) 1211 8.8.8.666 (1(1(18.8.8 1)
***,, innn mmHg NNNA 74...5 ((9.9) 828282.3.3 (((11.33) 80.22 (111.3) 7878.7. (1000.999)
*, iiinn n kgkgkg/m/m/m2 NANANA 272727..4.4 (((5.55 0)) 262626.6. (((4.44 5))) 22727.5 (((5.5.5.1)1)1 2666.5 ((555.2)2))
*, n (%) NANANA 1,020 (28.7%7%7%) 2,474747777 (44.5%) 1,1,1,143 (49.4%) 1,771 (27.8)
cholesterol*, in mg/dL NANANA 1212125.5.5.11 (3(334.4..5)5)) 131313444.888 (3(3(35.55 9)99 13131 8.8.8.2 2 2 (4(4(48.7) 137.3 (40.8)†
chchcholololesesesteteterororol*l*l , ininin mmmg/g/g/dLdLdL NANANA 505050 11.1 (((151515 11.1))) 545454 88.8 (((131313 55.5))) 585858 77.7 (((181818 99.9))) 585858 88.8 (((151515 99.9)†)†)†
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
23
Table 2. Significant Multiplicative Interaction Terms Between Genetic Risk Scores of Interest Associated with Ischemic Heart Disease Identified in the Meta-analyses.
Regression coefficient(Standard Error)
P-value P-value
Heterogeneity
GRSLDLxGRSTG
Hard events -0.047 (0.021) 0.027 0.011
All events -0.096 (0.028) 5.2x10-4 0.252
Non-pleiotropic GRSLDLxGRSIHD
Hard events 0.064 (0.022) 0.022 0.003
All events 0.091 (0.028) 1.2x10-3 0.461
GRS: Genetic risk score; IHD: Ischemic heart disease; TG: Triglycerides; LDL: Low-density lipoprotein.
Table 3: Results of the Improvement in Predictive Capacity when the GRSLDL and GRSTG, andthe Non-pleiotropic GRSLDL and GRSIHD Interaction Terms Were Added to the Model Based on the Individual Genetic Risk Scores: Changes in Discrimination C -Statistics) and in Reclassification (Continuous Net Reclassification Index –c-NRI– and Integrated Discrimination Improvement –IDI–) for the Two Ischemic Heart Disease Outcomes in the Meta-analyses.
Hard IHD Outcomes All IHD Outcomes
GRSLDL·GRSTG
c-statistic (p-value) 0.000 (0.471) 0.000 (0.217)
c-NRI (95% CI)* 0.011 (-0.030, 0.052) 0.030 (-0.021, 0.081)
IDI (95% CI)* 0.000 (-0.001, 0.001) 0.000 (-0.001, 0.002)
Non-pleiotropic GRSIHD·GRSLDL
c-statistic (p-value) 0.001 (0.263) 0.000 (0.637)
c-NRI (95% CI)* 0.031 (-0.011, 0.073) 0.029 (-0.019, 0.077)
IDI (95% CI)* 0.001 (-0.000, 0.001) 0.001 (-0.000, 0.003)
*c-NRI (95% CI): Continuous Net Reclassification Index (95% Confidence Interval); IDI (95% CI): Integrated Discrimination Improvement (95% Confidence Interval)
GRS: Genetic risk score; IHD: Ischemic heart disease; TG: Triglycerides; LDL: Low-densisisitytyty llipipipopopproror tetein.
TaTaTabblble 3: Resultsss ooof theee IIImprororovvevementntnt innn PPPredddiccctiveee CCCapppacacaciiityyy wwwhennn ttthe GGRRSLDDDLL and GGGRSRSRSTGGG, annnddhe e NoNoNon-n-n-plplpleieieioootrororopipipiccc GRGRGRSSSLDLL L aaandndnd GGGRSRR IHHHDDD InInInteteterararactcttioioion nn TTTerere mmmsss WeWeWereee AAAddddddededed tttoo o thhhee e MoMoModededel BaBaBasesesed d d ononn he Individual Genetetticicic Risk SSScococorer ss: Changeg s in DDDisii crc iminnnatatatioii n n C -Statistics) and in
Reclassification (Cooontntntinininuououoususus NNNetete RRRecececlalalasssssififificicicatattioioion nn IIIndndndeexex –––c-c-c-NRNRNRIII––– ananand d d InInIntetetegrgrgratatatede Discrimination mmprprp ovovememenent t –IDIDII–)–)) fofor r ththe e TTwowo IsIschchememicic HHeaeartrt DDisiseaeasese OuOutctcomomeses iin n ththe e MMetetaa-ananalalysysy eses..
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCGENETICS.115.001255
24
Figure Legends:
Figure 1: Forest Plot of the Association Between the Weighted Genetic Risk Scores for
Cardiovascular Risk Factors and Ischemic Heart Disease and the Prevalence/Incidence of Hard
Ischemic Heart Disease Events (Myocardial Infarction or Ischemic Heart Disease Death) Across
Studies and in the Meta-analysis.
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
Siscovick, Olle Melander, Krista Fischer, Veikko Salomaa and Jaume MarrugatSayols-Baixeras, Arto Pietilä, Maris Alver, Antonio Cabrera de León, Mariano Sentí, David
Roberto Elosua, Carla Lluís-Ganella, Isaac Subirana, Aki Havulinna, Kristi Läll, Gavin Lucas, SergiGreater than the Parts? A Genetic Approach
Cardiovascular Risk Factors and Ischemic Heart Disease: Is the Confluence of Risk Factors
Print ISSN: 1942-325X. Online ISSN: 1942-3268 Copyright © 2016 American Heart Association, Inc. All rights reserved.
TX 75231is published by the American Heart Association, 7272 Greenville Avenue, Dallas,Circulation: Cardiovascular Genetics
published online April 21, 2016;Circ Cardiovasc Genet.
http://circgenetics.ahajournals.org/content/early/2016/04/21/CIRCGENETICS.115.001255World Wide Web at:
The online version of this article, along with updated information and services, is located on the
http://circgenetics.ahajournals.org/content/suppl/2016/04/21/CIRCGENETICS.115.001255.DC1.htmlData Supplement (unedited) at:
http://circgenetics.ahajournals.org//subscriptions/
is online at: Circulation: Cardiovascular Genetics Information about subscribing to Subscriptions:
http://www.lww.com/reprints Information about reprints can be found online at: Reprints:
document. Permissions and Rights Question and Answer this process is available in the
located, click Request Permissions in the middle column of the Web page under Services. Further information aboutnot the Editorial Office. Once the online version of the published article for which permission is being requested is
can be obtained via RightsLink, a service of the Copyright Clearance Center,Circulation: Cardiovascular Genetics Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:
by guest on April 29, 2016http://circgenetics.ahajournals.org/Downloaded from
SUPPLEMENTAL MATERIAL
Description of the studies
Myocardial Infarction Genetics Consortium (MIGen)1: This case-control study included
2,967 cases of early-onset myocardial infarction (MI) (men ≤50 or women ≤60 years old),
diagnosed on the basis of autopsy evidence or a combination of chest pain,
electrocardiographic evidence, and elevation of cardiac biomarkers, and 3,075 age- and sex-
matched controls. Genome-wide genotype data and associated phenotype data were obtained
via the database of Genotypes and Phenotypes (dbGaP; http://dbgap.ncbi.nlm.nih.gov; project
number #5195). In this study, we performed a discovery analysis of GRS-GRS interactions on
myocardial infarction.
Framingham Heart Study (FHS)2: We included individuals from the FHS Offspring Cohort of
5,124 individuals that were offspring of the original participants and their spouses recruited in
1971. We selected participants aged 35 to 74 years at the time of the exam, who were free of
cardiovascular disease (CVD) at that time, and for whom DNA collected during the 1980s
and 1990s3 (3) and complete follow-up information was available. To maximize the number
of participants included in the analysis, we set exam 5 as the baseline visit for the Offspring
Cohort (3,557 individuals, 1991-95). All coronary events (myocardial infarction, angina,
coronary revascularization and death due to coronary artery disease) that occurred during
follow-up until the end of 2007 were considered as outcomes of interest. Genome-wide
genotype data and associated phenotype data for the FHS sample were also obtained via
dbGaP (project number #5195).
FINRISK (FINRISK 1997 & FINRISK 2002): The FINRISK cohorts comprise the respondents
of representative, cross-sectional population surveys that have been carried out every 5 years
since 1972 to assess the risk factors of chronic diseases and health behavior in the working
age population in 3-5 large study areas of Finland. DNA samples were collected in the
following survey years: 1987, 1992, 1997, 2002, 2007, and 2012. The MONICA and EHES
(EU) procedures were applied in phenotype collection and a wide spectrum of laboratory tests
was carried out from serum and plasma samples. Background information on socioeconomic
status, medical history, diet, exercise, anthropometric measures, etc. was collected by
questionnaires and during a clinical visit. The cohorts have been followed up by linking them
to the national hospital discharge register, causes-of-death register, and cancer register. The
study included 5699 individuals from the FINRISK 1997 cohort and 2426 from the FINRISK
2002 cohort.
Estonian Biobank (EGCUT)4: The Estonian Biobank is the population-based biobank of the
Estonian Genome Center at the University of Tartu (www.biobank.ee; EGCUT). The entire
project is conducted according to the Estonian Gene Research Act and all of the participants
have signed a broad informed consent. The cohort size is up to 50,750 individuals from 18
years of age and up, which closely reflects the age, sex, and geographical distribution of the
Estonian population. All of the subjects are recruited randomly by general practitioners and
physicians in hospitals. A Computer Assisted Personal interview is completed during 1 - 2
hours visit at a doctor’s office, which includes personal, genealogical, educational,
occupational history, and lifestyle data. Anthropometric measurements, blood pressure, and
resting heart rate are measured and venous blood taken during the visit. Medical history and
current health status is recorded according to ICD-10 codes.
The following ICD-9 and ICD-10 codes were initially used to identify potential coronary
events in the follow up of the cohorts:
a) Hard ischemic heart disease events (non-fatal and fatal MI and CHD deaths) using the
following codes:
i. ICD9: non fatal MI Code 410; CHD death Codes 410-414
ii. ICD10: non fatal MI Codes I21-l22; CHD death Codes I20-I22, I24.8, I24.9, I25
b) All ischemic heart disease events (additionally including angina and coronary
revascularization) using the following codes:
i. ICD9: non fatal CHD Codes 410-414; CHD death Codes 410-414
ii. ICD10: non fatal CHD Codes I20-22, I24.8, I24.9, I25; CHD death Codes I20-22,
I24.8, I24.9, I25.
Cases with a code of interest were further investigated and categorized by an event committee
in the Framingham Heart Study, FINRISK and the Estonian Biobank cohorts.
Statistical analysis: multiplicative and additive interactions
When the effect of one exposure on an outcome depends in some way on the presence or
absence of another exposure, there is interaction between the two exposures. Analysis of the
presence of interactions may be useful to identify which subgroups of the population would
benefit the most from an intervention when resources are limited, and also may provide
insight into the mechanisms for the outcome. Interactions can be assessed on different scales,
mainly additive or multiplicative, and both of them should be analysed and presented5.
Consider data on the table below showing the probability of an outcome according to the
exposure of two factors (A and B):
B=0 B=1
A=0 0.10 0.60
A=1 0.90 4.50
Where: p00 is the probability of the outcome when the individuals are not exposed to factor A
or B (p00=0.10), p10 is the probability of the individuals exposed to A but not to B (p10=0.90),
p01 is the probability of the individuals exposed to B but not to A (p01=0.60), and p11 is the
probability of presenting the outcome of those exposed both to A and B (p11=4.50).
A natural way to assess interaction is to measure the extent to which the effect of the
combination of the two factors together exceeds the effect of each considered individually:
[1] (p11 – p00) – [(p10 – p00) + (p01 – p00)] = p11 – p10 – p01 + p00
If this difference is nonzero, there is interaction on the additive scale. If the result is > 0, the
interaction is positive or “super-additive”; if the result is < 0, the interaction is negative or
“sub-additive”. For the data in Table 1, we can calculate: 4.50 –0.90 – 0.60 + 0.10 = 3.10, a
positive or “super-additive” interaction.
Sometimes, instead of using risk differences to measure effects, one might use risk ratios or
odds ratios. In our example: RR10 = 0.90 / 0.10 = 9; RR01 = 0.60 / 0.10 = 6; RR11 = 4.50 / 0.10
= 45.
A measure of interaction on the multiplicative scale for risk ratios could then be taken as:
[2] RR11 / (RR10 · RR01)
measuring the extent to which, on the risk ratio scale, the effect of both exposures together
exceeds the product of the effects of the two exposures considered separately. If the ratio is 1,
then the effect of both exposures together is equal to the product of the effect of the two
exposures considered separately, that is, there is no interaction on the multiplicative scale. If
the ratio is > 1, the multiplicative interaction is positive. If the ratio is < 1, the multiplicative
interaction is negative.
In our example: 45 / (9 · 6) = 45 / 54 = 0.83, the multiplicative interaction is negative on the
risk ratio scale. In general, measures of multiplicative interaction on the odds ratio and risk
ratio scales will be very close to one another whenever the outcome is rare.
We may assess additive interaction from data when only relative risks (or odds ratios) are
reported. If we divide eq. [1] by p00 we obtain the following:
[3] p11/p00 – p10/p00 – p01/p00 + p00/p00 = RR11 – RR10 – RR01 + 1
This quantity is referred to as the “relative excess risk due to interaction” or RERI6. We will
assume a super-additive interaction if RERI > 0, a sub-additive interaction if RERI < 0, and
an absence of interaction on the additive scale if RERI = 0.
Another commonly used metric to assess the presence of an additive interaction using relative
risk values is the “synergy index” that measures the extent to which the risk ratio for both
exposures together exceeds 1, and whether this is greater than the sum of the extent to which
each of the risk ratios, considered separately, exceeds 1:
[4] Synergy Index = (RR11 – 1) / [(RR10 – 1) + (RR01 – 1)]
We will assume a super-additive interaction if this index > 1, a sub-additive interaction if the
index < 1, and an absence of interaction on the additive scale if the index = 1.
What differential information do we obtain from the additive and the multiplicative
interaction? The additive scale is useful for assessing the public health importance of
interventions and the public
health significance of interaction and for targeting subpopulations for which the intervention
is most effective. A second reason sometimes given for using additive interaction is that it
more closely corresponds to tests for mechanistic interaction. Finally, tests for additive
interaction usually are more powerful than tests for multiplicative interaction.
On the other hand, the multiplicative scale is the most natural scale on which to assess
interaction for logistic or survival regression models. Second, some authors claim that there
is, in general, less heterogeneity on the multiplicative than on the additive scale. Another
reason to assess multiplicative interaction is that the relative effect measures are better suited
to assessing causality, although the question whether the relative or absolute measure is more
useful for assessing causality varies by setting7.
Supplementary Table 1. Genetic variants identified to be robustly associated with cardiovascular risk factors or coronary endpoints and included in the different genetic risk scores. The associated gene, possible alleles, risk allele, risk allele frequency (RAF), p-value and effect size of the association between each variant and the specific trait of interest, risk allele frequency in the different studies, and references supporting the selection of the genetic variants. Ischemic heart disease (IHD)
Top SNP Gene Chr Risk allele RAF p-value OR References
rs11206510 PCSK9 1 T 0.85 2.34E-08 1.08 (8,9,10)
rs17114036 PPAP2B 1 A 0.92 2.22E-13 1.13 (8,10)
rs646776 CELSR2-PSRC1-SORT1 1 T 0.75 9.01E-19 1.11 (9)
rs4845625 IL6R 1 T 0.45 3.93E-08 1.05 (10)
rs17464857 MIA3 1 T 0.86 4.18E-05 1.06 (10)
rs17465637 MIA3 1 C 0.66 3.52E-12 1.08 (8,9)
rs515135 APOB 2 C 0.79 3.09E-08 1.07 (10)
rs6544713 ABCG5-ABCG8 2 T 0.32 8.88E-07 1.05 (8,10)
rs1561198 VAMP5-VAMP8-GGCX 2 T 0.46 6.37E-10 1.06 (10)
rs2252641 ZEB2-ACO74093.1 2 C 0.48 5.16E-04 1.03 (9,10)
rs6725887 WDR12 2 C 0.11 9.51E-18 1.14 (8,10)
rs9818870 MRAS 3 T 0.14 2.21E-06 1.07 (9,10)
rs1878406 EDNRA 4 T 0.16 1.24E-06 1.06 (10)
rs7692387 GUCY1A3 4 G 0.81 7.35E-09 1.07 (10)
rs273909 SLC22A4-SLC22A5 5 G 0.12 1.24E-04 1.06 (10)
rs12526453 PHACTR1 6 C 0.71 2.14E-20 1.10 (8,9,10)
rs17609940 ANKSIA 6 G 0.82 3.00E-02 1.03 (8,10)
rs10947789 KCKN5 6 T 0.78 1.63E-06 1.05 (10)
rs12190287 TCF21 6 C 0.62 1.07E-03 1.06 (8,10)
rs2048327 SLC22A3-LPAL2-LPA 6 C 0.35 2.46E-09 1.06 (10)
rs3798220 LPA 6 C 0.02 4.66E-09 1.42 (8,10)
rs4252120 PLG 6 T 0.74 3.32E-03 1.03 (10)
rs2023938 HDAC9 7 C 0.10 1.36E-04 1.06 (10)
rs10953541 BCAP29 7 C 0.78 1.02E-05 1.05 (9)
rs11556924 ZC3HC1 7 C 0.69 5.34E-11 1.08 (8,10)
rs264 LPL 8 G 0.85 1.06E-05 1.06 (10)
rs2954029 TRIB1 8 A 0.55 2.61E-06 1.04 (8,10)
rs4977574 CDKN2A, CDKN2B 9 G 0.49 6.35E-98 1.21 (8,9)
rs579459 ABO 9 C 0.21 1.14E-10 1.08 (8,10)
rs2505083 KIAA1462 10 C 0.40 1.57E-10 1.06 (10)
rs2047009 CXCL12 10 G 0.48 2.75E-11 1.06 (10)
rs501120 CXCL12 10 T 0.81 1.39E-11 1.08 (10)
rs11203042 LIPA 10 T 0.45 1.22E-04 1.04 (10)
rs1412444 LIPA 10 T 0.37 5.15E-12 1.07 (10)
rs12413409 CYP17A1-CNM2-NT5C2 10 G 0.89 1.07E-07 1.08 (8,10)
rs974819 PDGFD 11 T 0.33 2.44E-10 1.07 (10)
rs964184 ZNF259, APOA5- 11 G 0.18 5.60E-05 1.05 (8)
A4-C3-A1
rs7136259 ATP2B1 12 T 0.43 2.45E-05 1.04 (9)
rs3184504 SH2B3 12 T 0.42 1.03E-09 1.07 (8,9,10)
rs9319428 FLT1 13 A 0.31 7.13E-05 1.04 (10)
rs4773144 COLA1-COLA2 13 G 0.43 3.87E-07 1.05 (8,10)
rs9515203 COLA1-COLA2 13 T 0.76 9.33E-10 1.07 (10)
rs2895811 HHIPL1 14 C 0.41 1.86E-05 1.04 (8,10)
rs7173743 ADAMTS7 15 T 0.56 5.55E-16 1.08 (10)
rs17514846 FURIN-FES 15 A 0.44 3.10E-07 1.05 (10)
rs216172 SMG6 17 C 0.35 5.07E-07 1.05 (8,10)
rs12936587 RAI1-PEMT-RASD1 17 G 0.61 8.24E-04 1.03 (8,10)
rs46522 UBESZ 17 T 0.51 1.84E-05 1.04 (8,10)
rs1122608 LDLR 19 G 0.77 2.73E-11 1.08 (8,9,10)
rs2075650 ApoE-ApoC1 19 G 0.13 1.61E-06 1.07 (10)
rs445925 ApoE-ApoC1 19 G 0.90 4.23E-06 1.09 (10)
rs9982601 Gene desert (KCNE2) 21 T 0.13 1.33E-13 1.12 (8,9,10)
rs17087335 REST-NOA1 4 T 0.21 4.60E-08 1.06 (11)
rs3918226 NOS3 7 T 0.06 1.70E-09 1.14 (11)
rs10840293 SWAP70 11 A 0.55 1.30E-08 1.06 (11)
rs56062135 SMAD3 15 C 0.79 4.50E-09 1.07 (11)
rs8042271 MFGE8-ABHD2 15 G 0.93 3.70E-08 1.10 (11)
rs7212798 BCAS3 17 C 0.15 1.90E-08 1.08 (11)
rs663129 PMAIP1-MC4R 18 A 0.26 3.20E-08 1.06 (11)
rs180803 POM12IL9P-ADORA2A 22 G 0.97 1.60E-10 1.20 (11)
LDL
Top SNP Gene Chr Risk allele
RAF p-value Effect size (mg/dL) Reference
rs12027135 LDLRAP1 1 T 0.53 4x10-11 1.1 (12) rs2479409 PCSK9 1 G 0.3 2X10-28 2.01 (12) rs3850634 ANGPTL3 1 T 0.68 5x10-41 1.59 (12) rs629301 SORT1 1 T 0.78 1x10-170 5.65 (12)
rs2807834 MOSC1 1 G 0.68 6x10-11 1.09 (12) rs514230 IRF2BP2 1 T 0.52 9x10-12 1.13 (12)
rs1367117 APOB 2 A 0.3 4x10-114 4.05 (12) rs4299376 ABCG5/8 2 G 0.3 2x10-47 2.75 (12)
rs12916 HMGCR 5 C 0.39 5x10-45 2.45 (12) rs6882076 TIMD4 5 C 0.65 2x10-22 1.67 (12) rs3757354 MYLIP 6 C 0.78 1x10-11 1.43 (12) rs1800562 HFE 6 G 0.94 6x10-10 2.22 (12) rs3177928 HLA 6 A 0.16 2x10-15 1.83 (12)
rs11153594 FRK 6 C 0.59 3x10-9 0.89 (12) rs1564348 LPA 6 C 0.17 2x10-17 1.95 (12)
rs12670798 DNAH11 7 C 0.23 7x10-10 1.26 (12) rs217386 NPC1L1 7 G 0.57 4x10-11 1.17 (12)
rs2126259 PPP1R3B 8 C 0.9 7x10-15 2.22 (12)
rs1030431 CYP7A1 8 A 0.35 4x10-9 0.95 (12) rs2954022 TRIB1 8 C 0.54 3x10-29 1.84 (12)
rs11136341 PLEC1 8 G 0.4 4x10-13 1.4 (12) rs649129 ABO 9 T 0.22 8x10-22 2.05 (12)
rs1129555 GPAM 10 A 0.29 2x10-9 1.08 (12) rs174583 FADS1-2-3 11 C 0.65 1x10-21 1.71 (12) rs964184 APOA1-C3-
A4-A5 11 G 0.13 1x10-26 2.85 (12)
rs11220462 ST3GAL4 11 A 0.14 1x10-15 1.95 (12) rs11065987 BRAP 12 A 0.58 1x10-9 0.97 (12) rs1169288 HNF1A 12 C 0.33 1x10-15 1.42 (12) rs2332328 NYNRIN 14 T 0.48 4x10-11 1.17 (12) rs247616 CETP 16 C 0.68 9x10-13 1.45 (12)
rs2000999 HPR 16 A 0.2 2x10-22 2 (12) rs7225700 OSBPL7 17 C 0.65 4x10-9 0.87 (12) rs6511720 LDLR 19 G 0.89 4x10-117 6.99 (12)
rs10401969 CILP2 19 T 0.93 7x10-22 3.11 (12) rs4420638 APOE-C1-C2 19 G 0.17 9x10-147 7.14 (12) rs2902941 MAFB 20 A 0.67 1x10-8 0.98 (12)
rs909802 TOP1 20 T 0.47 3x10-19 1.41 (12)
rs267733 ANXA9-CERS2
1 A 0.84 5x10-9 0.87
(13)
rs12748152 PIGV-NROB2
1 T 0.09 3X10-12 0.87
(13)
rs2710642 EHBP1 2 A 0.65 6x10-9 0.87
(13)
rs10490626 INSIG2 2 G 0.92 2x10-12 0.87
(13)
rs2030746 LOC84931 2 T 0.4 9x10-9 0.87
(13)
rs1250229 FN1 2 C 0.73 3x10-8 0.87
(13)
rs11563251 UGT1A1 2 T 0.12 5x10-8 0.87
(13)
rs7640978 CMTM6 3 C 0.91 1x10-8 0.87
(13)
rs17404153 ACAD11 3 G 0.86 2x10-9 0.87
(13)
rs6831256 LRPAP1 4 G 0.42 2x10-8 0.87
(13)
rs4530754 CSNK1G3 5 A 0.54 4x10-12 0.87
(13)
rs4722551 MIR148A 7 C 0.2 4x10-14 0.87
(13)
rs10102164 SOX17 8 A 0.21 4x10-11 0.87
(13)
rs3780181 VLDLR 9 A 0.92 2x10-9 0.87
(13)
rs4942486 BRCA2 13 T 0.48 2x10-11 0.87
(13)
rs1801689 APOH-PRXCA
17 C 0.04 1x10-11 0.87
(13)
rs314253 DLG4 17 T 0.63 3x10-10 0.87
(13)
rs364585 SPTLC3 20 G 0.62 4x10-10 0.87
(13)
rs2328223 SNX5 20 C 0.21 6x10-9 0.87
(13)
rs5763662 MTMR3 22 T 0.04 1x10-8 0.87
(13)
rs4253772 PPARA 22 T 0.11 3x10-8 0.87
(13)
HDL
Top SNP Gene Chr Risk allele RAF p-value Effect size (mg/dL) References
rs1042034 APOB 2 T 0.78 1x10-30 -0.9 (12)
rs10808546 TRIB1 8 C 0.56 6x10-19 -0.61 (12)
rs12328675 COBLL1 2 T 0.87 3x10-10 -0.68 (12)
rs12678919 LPL 8 A 0.88 9x10-98 -2.25 (12)
rs12967135 MC4R 18 A 0.23 7x10-09 -0.42 (12)
rs13107325 SLC39A8 4 T 0.07 7x10-11 -0.84 (12)
rs1515100 IRS1 2 A 0.63 2x10-09 -0.46 (12)
rs1532085 LIPC 15 G 0.61 3x10-96 -1.45 (12)
rs1689800 ZNF648 1 G 0.35 3x10-10 -0.47 (12)
rs16942887 LCAT 16 G 0.88 8x10-33 -1.27 (12)
rs17145738 MLXIPL 7 C 0.88 1x10-09 -0.57 (12)
rs174601 FADS1-2-3 11 T 0.36 2x10-22 -0.73 (12)
rs181362 UBE2L3 22 T 0.2 1x10-08 -0.46 (12)
rs1883025 ABCA1 9 T 0.25 2x10-33 -0.94 (12)
rs2293889 TRPS1 8 T 0.41 6x10-11 -0.44 (12)
rs2652834 LACTB 15 A 0.2 9x10-09 -0.39 (12)
rs2814944 C6orf106 6 A 0.16 4x10-09 -0.49 (12)
rs2923084 AMPD3 11 G 0.17 5x10-08 -0.41 (12)
rs2925979 CMIP 16 T 0.3 2x10-11 -0.45 (12)
rs3136441 LRP4 11 T 0.85 3x10-18 -0.78 (12)
rs3741414 LRP1 12 C 0.76 2x10-08 -0.46 (12)
rs3764261 CETP 16 C 0.68 7x10-380 -3.39 (12)
rs4082919 PGS1 17 G 0.48 5x10-09 -0.4 (12)
rs4148008 ABCA8 17 G 0.32 2x10-10 -0.42 (12)
rs4420638 APOE-C1-C2 19 G 0.17 4x10-21 -1.06 (12)
rs4660293 PABPC4 1 G 0.23 4x10-10 -0.48 (12)
rs4731702 KLF14 7 C 0.52 1x10-15 -0.59 (12)
rs4759375 SBNO1 12 C 0.94 8x10-09 -0.86 (12)
rs4765127 ZNF664 12 G 0.66 3x10-10 -0.44 (12)
rs4846914 GALNT2 1 G 0.4 4x10-21 -0.61 (12)
rs605066 CITED2 6 C 0.42 3x10-08 -0.39 (12)
rs6065906 PLTP 20 C 0.18 2x10-22 -0.93 (12)
rs643531 TTC39B 9 C 0.14 1x10-13 -0.72 (12)
rs6450176 ARL15 5 A 0.26 5x10-08 -0.49 (12)
rs7115089 UBASH3B 11 C 0.63 3x10-08 -0.31 (12)
rs7134375 PDE3A 12 C 0.58 4x10-08 -0.4 (12)
rs7134594 MVK 12 C 0.47 7x10-15 -0.44 (12)
rs7241918 LIPG 18 G 0.17 3x10-49 -1.31 (12)
rs7255436 ANGPTL4 19 C 0.47 3x10-08 -0.45 (12)
rs737337 LOC55908 19 C 0.08 3x10-09 -0.64 (12)
rs838880 SCARB1 12 T 0.69 3x10-14 -0.61 (12)
rs881844 STARD3 17 C 0.34 3x10-14 -0.51 (12)
rs964184 APOA1-C3-A4-A5 11 G 0.13 5x10-47 -1.5 (12)
rs9987289 PPP1R3B 8 A 0.09 6x10-25 -1.21 (12)
rs1084651 LPA 6 A 0.16 3X10-8 -0.56 (12)
rs386000 LILRA 8 G 0.8 4X10-16 -0.83 (12)
rs1800961 HNF4A 20 T 0.03 1X10-15 -1.88 (12)
rs12748152 PIGV- 1 T 0.09 1x10-15 -0.31 (13)
NROB2
rs12145743 HDGF-PMVK 1 T 0.66 2X10-8 -0.31 (13)
rs4650994 ANGPTL1 1 A 0.51 7x10-9 -0.31 (13)
rs1047891 CPS1 2 A 0.33 9x10-10 -0.31 (13)
rs2606736 ATG7 3 T 0.61 5x10-8 -0.31 (13)
rs2290547 SETD2 3 A 0.2 4x10-9 -0.31 (13)
rs2013208 RBM5 3 C 0.5 9x10-12 -0.31 (13)
rs13326165 STAB1 3 G 0.79 9x1011 -0.31 (13)
rs6805251 GSK3B 3 C 0.61 1x10-8 -0.31 (13)
rs17404153 ACAD11 3 T 0.14 5x10-9 -0.31 (13)
rs10019888 C4orf52 4 G 0.18 5x10-8 -0.31 (13)
rs3822072 FAM13A 4 A 0.46 4x10-12 -0.31 (13)
rs2602836 ADH5 4 G 0.56 5x10-8 -0.31 (13)
rs1936800 RSPO3 6 T 0.51 3x10-10 -0.31 (13)
rs998584 VEGFA 6 A 0.49 2x10-11 -0.31 (13)
rs702485 DAGLB 7 A 0.55 6x10-12 -0.31 (13)
rs4142995 SNX13 7 T 0.38 9x10-12 -0.31 (13)
rs4917014 IKZF1 7 T 0.68 1x10-8 -0.31 (13)
rs17173637 TMEM176A 7 C 0.12 2x10-8 -0.31 (13)
rs970548 MARCH8-ALOX5 10 A 0.74 2x10-10 -0.31 (13)
rs11246602 OR4C46 11 A 0.85 2x10-10 -0.31 (13)
rs12801636 KAT5 11 G 0.77 3x10-8 -0.31 (13)
rs499974 MOGAT2-DGAT2 11 A 0.19 1x10-8 -0.31 (13)
rs4983559 ZBTB42-AKT1 14 A 0.6 1x10-8 -0.31 (13)
rs1121980 FTO 16 A 0.43 7x10-9 -0.31 (13)
rs17695224 HAS1 19 A 0.26 2x10-13 -0.31 (13)
rs731839 PEPD 19 G 0.35 3x10-9 -0.31 (13)
TG
Top SNP Gene Chr Risk allele RAF p-value Effect size (mg/dL) Reference
rs10195252 COBLL1 2 T 0.60 1.63x10-10 2.01 (12)
rs10401969 CILP2 19 T 0.93 1.61x10-29 7.83 (12)
rs1042034 APOB 2 T 0.78 1.36x10-45 5.99 (12)
rs10761731 JMJD1C 10 A 0.57 3.48x10-12 2.38 (12)
rs11613352 LRP1 12 C 0.77 4.43x10-10 2.70 (12)
rs11649653 CTF1 16 C 0.60 3.35x10-08 2.13 (12)
rs11776767 PINX1 8 C 0.37 1.30x10-08 2.01 (12)
rs12310367 ZNF664 12 A 0.66 1.21x10-08 2.42 (12)
rs1260326 GCKR 2 T 0.41 6.00x10-133 8.76 (12)
rs12678919 LPL 8 A 0.88 1.50x10-115 13.64 (12)
rs1321257 GALNT2 1 G 0.39 2.09x10-14 2.76 (12)
rs1495743 NAT2 8 G 0.22 4.11x10-14 2.97 (12)
rs1553318 TIMD4 5 C 0.64 3.68x10-12 2.63 (12)
rs174546 FADS1-2-3 11 T 0.34 5.41x10-24 3.82 (12)
rs2068888 CYP26A1 10 G 0.53 2.38x10-08 2.28 (12)
rs2131925 ANGPTL3 1 T 0.68 8.84x10-43 4.94 (12)
rs2412710 CAPN3 15 A 0.02 1.87x10-08 7.00 (12)
rs261342 LIPC 15 G 0.22 2.42x10-13 2.99 (12)
rs2929282 FRMD5 15 T 0.05 1.63x10-11 5.13 (12)
rs2943645 IRS1 2 T 0.63 2.35x10-08 1.89 (12)
rs2954029 TRIB1 8 A 0.53 3.29x10-55 5.64 (12)
rs439401 APOE-C1-C2 19 C 0.64 1.14x10-30 5.50 (12)
rs442177 KLHL8 4 T 0.59 8.65x10-12 2.25 (12)
rs4810479 PLTP 20 C 0.24 4.69x10-18 3.32 (12)
rs5756931 PLA2G6 22 T 0.60 3.82x10-08 1.54 (12)
rs645040 MSL2L1 3 T 0.78 2.52x10-08 2.22 (12)
rs7205804 CETP 16 G 0.55 1.00x10-12 2.88 (12)
rs7811265 MLXIPL 7 A 0.81 9.00x10-43 7.91 (12)
rs964184 APOA1-C3-A4-A5 11 G 0.13 7.00x10-240 16.95 (12)
rs9686661 MAP3K1 5 T 0.20 1.32x10-10 2.57 (12)
rs2247056 HLA 6 C 0.75 1X10-15 2.99 (12)
rs13238203 TYW1B 7 C 0.96 1X10-9 7.91 (12)
rs12748152 PIGV-NROB2 1 T 0.09 1x10-9 1.54 (13)
rs6831256 LRPAP1 4 G 0.42 2x10-12 1.54 (13)
rs998584 VEGFA 6 A 0.49 3x10-15 1.54 (13)
rs1936800 RSPO3 6 T 0.51 3x10-8 1.54 (13)
rs38855 MET 7 A 0.53 2x10-8 1.54 (13)
rs4722551 MIR148A 7 C 0.2 9x10-11 1.54 (13)
rs1832007 AKR1CA 10 A 0.82 2x10-12 1.54 (13)
rs3198697 PDXDC1 16 C 0.57 2x10-8 1.54 (13)
rs1121980 FTO 16 A 0.43 3x10-8 1.54 (13)
rs8077889 MPP3 17 C 0.22 1x10-8 1.54 (13)
rs7248104 INSR 19 G 0.58 5x10-10 1.54 (13)
rs731839 PEPD 19 G 0.35 3x10-9 1.54 (13)
BP
Top SNP Gene Chr Risk allele RAF p-value Effect size (mm Hg) Standardized β Reference
rs10850411 TBX5–TBX3 12 T 0.70 5.40x10-08 0.35 0.017 (14)
rs11222084* ADAMTS8 11 T 0.38 4.00x10-04 0.26 0.013 (15)
rs12940887 ZNF652 17 T 0.38 1.80x10-10 0.36 0.018 (14)
rs13002573* FIGN 2 A 0.80 3.25x10-07 0.42 0.020 (15)
rs13082711* TBX5-TBX3 12 C 0.22 1.50x10-06 0.32 0.016 (14)
rs13107325 SLC39A8 4 C 0.95 3.30x10-14 0.98 0.048 (14)
rs13139571* GUCY1A3-GUCY1B3 4 C 0.76 1.20x10-06 0.32 0.016 (14)
rs1327235 JAG1 20 G 0.46 1.90x10-08 0.34 0.016 (14)
rs1378942 CYP1A1-ULK3 15 C 0.35 5.70x10-23 0.61 0.030 (14)
rs1446468 FIGN 2 C 0.47 1.82x10-12 0.50 0.024 (15)
rs17367504 MTHFR-NPPB 1 A 0.85 8.70x10-22 0.90 0.044 (14)
rs17477177 PIK3CG 7 C 0.28 5.67x10-11 0.55 0.026 (15)
rs17608766 GOSR2 17 C 0.14 1.10x10-10 0.56 0.027 (14)
rs1813353 CACNB2 10 T 0.68 2.60x10-12 0.57 0.028 (14)
rs2071518* NOV 8 T 0.17 2.08x10-02 0.18 0.009 (15)
rs2521501 FURIN-FES 15 A 0.31 5.20x10-19 0.65 0.032 (14)
rs2782980* ADRB1 10 C 0.80 7.66x10-07 0.41 0.019 (15)
rs2932538 MOV10 1 G 0.75 1.20x10-09 0.39 0.019 (14)
rs319690 MAP4 3 T 0.51 4.74x10-08 0.42 0.020 (15)
rs3774372* ULK4 3 C 0.17 3.90x10-01 0.07 0.003 (14)
rs381815 PLEKHA7 11 T 0.26 5.30x10-11 0.58 0.028 (14)
rs419076 MECOM 3 T 0.47 1.80x10-13 0.41 0.020 (14)
rs4373814 CACNB2 10 C 0.45 4.80x10-11 0.37 0.018 (14)
rs4590817 C10orf107 10 G 0.84 4.00x10-12 0.65 0.032 (14)
rs633185 FLJ32810-TMEM133 11 G 0.72 1.20x10-17 0.57 0.028 (14)
rs7129220 ADM 11 A 0.11 3.00x10-12 0.62 0.031 (14)
rs871606* CHIC2 4 T 0.85 3.04x10-04 0.40 0.019 (15)
rs932764 PLCE1 10 G 0.44 7.10x10-16 0.48 0.024 (14)
Diabetes
Top SNP Gene Chr Risk allele RAF p-value OR Reference
rs10811661 CDKN2A/B 9 T 0.83 1.45x10-10 1.19 (16,17,18)
rs11634397 ZFAND6 15 G 0.4 2.40x10-09 1.06 (16)
rs13292136 CHCHD9 9 C 0.93 2.80x10-08 1.11 (16)
rs1387153 MTNR1B 11 T 0.28 7.80x10-15 1.09 (16,17)
rs1470579 IGF2BP2 3 C 0.31 2.17x10-09 1.14 (17,18)
rs1531343 HMGA2 12 C 0.1 3.60x10-09 1.1 (16)
rs1552224 CENTD2 11 A 0.88 1.40x10-22 1.14 (16)
rs231362 KCNQ1 11 G 0.52 2.80x10-13 1.08 (16)
rs243021 BCL11A 2 A 0.46 2.90x10-15 1.08 (16)
rs3802177 SLC30A8 8 G 0.66 1.45x10-08 1.15 (17,18)
rs4457053 ZBED3 5 G 0.26 2.80x10-12 1.08 (16)
rs5015480 HHEX/IDE 10 C 0.58 1.33x10-15 1.18 (17,18)
rs7578326 IRS1 2 A 0.64 5.40x10-20 1.11 (16)
rs10440833 CDKAL1 6 A 0.26 2.00x10-22 1.25 (17,18)
rs7903146 TCF7L2 10 T 0.22 2.40x10-518 1.4 (17,18)
rs7957197 HNF1A 12 T 0.85 2.40x10-10 1.07 (16)
rs8042680 PRC1 15 A 0.74 2.40x10-10 1.07 (16)
rs896854 TP53INP1 8 T 0.48 9.90x10-10 1.06 (16)
rs972283 KLF14 7 G 0.55 2.20x10-10 1.07 (16)
rs849134 JAZF1 7 A 0.7 3X10-9 1.13 (17)
rs11642841 FTO 16 A 0.18 3x10-8 1.13 (17)
rs1801214 WFS1 4 T 0.27 3x10-8 1.13 (16)
rs5219 KCNJ11 11 T 0.26 6.7x10-11 1.14 (16)
BMI
Top SNP Gene Chr Risk allele RAF p-value Effect size(Kg/m2) Reference
rs657452 AGBL4 1 A 0.39 5.48x10-13 0.023 (19)
rs12286929 CADM1 11 G 0.52 1.31x10-12 0.022 (19)
rs7903146 TCF7L2 10 C 0.71 1.11x10-11 0.023 (19)
rs10132280 STXBP6 14 C 0.68 1.14x10-11 0.023 (19)
rs17094222 HIF1AN 10 C 0.21 5.94x10-11 0.025 (19)
rs7599312 ERBB4 2 G 0.72 1.17x10-10 0.022 (19)
rs2365389 FHIT 3 C 0.58 1.63x10-10 0.02 (19)
rs2820292 NAV1 1 C 0.56 1.83x10-10 0.02 (19)
rs12885454 PRKD1 14 C 0.64 1.94x10-10 0.021 (19)
rs16851483 RASA2 3 T 0.07 3.55x10-10 0.048 (19)
rs1167827
HIP1; PMS2L3; PMS2P5; WBSCR16
7 G 0.55 6.33x10-10 0.02
(19)
rs758747 NLRC3 16 T 0.27 7.47x10-10 0.023 (19)
rs1928295 TLR4 9 T 0.55 7.91x10-10 0.019 (19)
rs9925964 KAT8;ZNF646; VKORC1 16 A 0.62 8.11x10-10 0.019 (19)
rs11126666 KCNK3 2 A 0.28 1.33x10-9 0.021 (19)
rs2650492 SBK1; APOBR 16 A 0.30 1.92x10-9 0.021 (19)
rs6804842 RARB 3 G 0.58 2.48x10-9 0.019 (19)
rs4740619 C9orf93 9 T 0.54 4.56x10-9 0.018 (19)
rs13191362 PARK2 6 A 0.88 7.34x10-9 0.028 (19)
rs3736485 SCG3; DMXL2 15 A 0.45 7.41x10-9 0.018 (19)
rs17001654 NUP54; SCARB2 4 G 0.15 7.76x10-9 0.031 (19)
rs11191560 NT5C2;
CYP17A1; SFXN2
10 C 0.09 8.45x10-9 0.031 (19)
rs1528435 UBE2E3 2 T 0.63 1.20x10-8 0.018 (19)
rs1000940 RABEP1 17 G 0.32 1.28C10-8 0.019 (19)
rs2033529 TDRG1; LRFN2 6 G 0.29 1.39x10-8 0.019 (19)
rs11583200 ELAVL4 1 C 0.40 1.48x10-8 0.018 (19)
rs9400239 FOXO3; HSS00296402 6 C 0.69 1.61x10-8 0.019 (19)
rs10733682 LMX1B 9 A 0.48 1.83x10-8 0.017 (19)
rs11688816 EHBP1 2 G 0.56 1.89x10-8 0.017 (19)
rs11057405 CLIP1 12 G 0.90 2.02x10-8 0.031 (19)
rs11727676 HHIP 4 T 0.91 2.55x10-8 0.036 (19)
rs3849570 GBE1 3 A 0.36 2.60x10-8 0.019 (19)
rs6477694 EPB41L4B; C9orf4 9 C 0.37 2.67x10-8 0.017 (19)
rs7899106 GRID1 10 G 0.05 2.96x10-8 0.04 (19)
rs2176598 HSD17B12 11 T 0.25 2.97 x10-8 0.02 (19)
rs2245368 PMS2L11 7 C 0.18 3.19 x10-8 0.032 (19)
rs17724992 GDF15; PGPEP1 19 A 0.75 3.42 x10-8 0.019 (19)
rs7243357 GRP 18 T 0.81 3.86 x10-8 0.022 (19)
rs2033732 RALYL 8 C 0.75 4.89x10-8 0.019 (19)
rs1558902 FTO 16 A 0.42 7.51x10-153 0.082 (19)
rs6567160 MC4R 18 C 0.24 3.93x10-53 0.056 (19)
rs13021737 TMEM18 2 G 0.83 1.11x10-50 0.06 (19)
rs10938397 GNPDA2; GABRG1 4 G 0.43 3.21x10-38 0.04 (19)
rs543874 SEC16B 1 G 0.19 2.62x10-35 0.048 (19)
rs2207139 TFAP2B 6 G 0.18 4.13x10-29 0.045 (19)
rs11030104 BDNF 11 A 0.79 5.56x10-28 0.041 (19)
rs3101336 NEGR1 1 C 0.61 2.66x10-26 0.033 (19)
rs7138803 BCDIN3D; FAIM2 12 A 0.38 8.15x10-24 0.032 (19)
rs10182181 ADCY3; POMC; NCOA1
2 G 0.46 8.78x10-24 0.031 (19)
rs3888190
SH281; APOBR; ATXN2L;
SBK1; SULT1A2;
TUFM
16 A 0.40 3.14x10-23 0.031
(19)
rs1516725 ETV5 3 C 0.87 1.89x10-22 0.045 (19)
rs12446632 GPRC5B; IQCK 16 G 0.87 1.48x10-18 0.04 (19)
rs2287019 QPCTL; G/PR 19 C 0.80 4.59x10-18 0.036 (19)
rs16951275 MAP2K5; LBXCOR1 15 T 0.78 1.91x10-17 0.031 (19)
rs3817334 MTCH2;
C1QTNF4; SPI1; CELF1
11 T 0.41 5.15x10-17 0.026 (19)
rs2112347 POC5;
HMGCR; COL4A3BP
5 T 0.63 6.19x10-17 0.026 (19)
rs12566985 FPGT-TNN/3K 1 G 0.45 3.28x10-15 0.024 (19)
rs3810291 ZC3H4 19 A 0.67 4.81x10-15 0.028 (19)
rs7141420 NRXN3 14 T 0.53 1.23x10-14 0.024 (19)
rs13078960 CADM2 3 G 0.20 1.74x10-14 0.03 (19)
rs10968576 LING02 9 G 0.32 6.61x10-14 0.025 (19)
rs17024393 GNAT2; AMPD2 1 C 0.04 7.03x10-14 0.066 (19)
rs12429545 OLFM4 13 A 0.13 1.09x10-12 0.033 (19)
rs13107325 SLC39A8 4 T 0.07 1.83x10-12 0.048 (19)
rs11165643 PTBP2 1 T 0.58 2.07x10-12 0.022 (19)
rs17405819 HNF4G 8 T 0.70 2.07x10-11 0.022 (19)
rs1016287 LINC01122 2 T 0.29 2.25x10-11 0.023 (19)
rs4256980 TR/M66; TUB 11 G 0.65 2.90x10-11 0.021 (19)
rs12401738 FUBP1; USP33 1 A 0.35 1.15x10-10 0.021 (19)
rs205262 C6orf106; SNRPC 6 G 0.27 1.75x10-10 0.022 (19)
rs9581854 MTIF3; GTF3A 13 T 0.20 2.29x10-10 0.03 (19)
rs12940622 RPTOR 17 G 0.58 2.49x10-09 0.018 (19)
rs11847697 PRKD1 14 T 0.04 3.99x10-09 0.049 (19)
rs2075650 TOMM40; APOE; APOC1 19 A 0.85 1.25x10-08 0.026 (19)
rs2121279 LRP1B 2 T 0.15 2.31 x10-08 0.025 (19)
rs29941 KCTD15 19 G 0.67 2.41 x10-08 0.018 (19)
rs1808579 NPC1; C18orf8 18 C 0.53 4.17 x10-08 0.017 (19)
rs9641123 CALCR; hsa-miR-653 7 C 0.43 2.80x10-10 0.029 (19)
rs4787491 MAPK3; KCTD13;
INO80E; … 16 G 0.51 2.70x10-8 0.022
(19)
rs9540493 MIR548X2; PCDH9 13 A 0.45 4.97x10-8 0.021 (19)
rs9374842 LOC285762 6 T 0.74 2.67x10-8 0.023 (19)
Waist
Top SNP Gene Chr Risk allele RAF p-value Effect size (mg/dl) References
rs905938 DCST2 1 T 0.74 7.30E-10 0.025 (20)
rs10919388 GORAB 1 C 0.72 3.20E-09 0.024 (20)
rs1385167 MEIS1 2 G 0.15 1.90E-09 0.029 (20)
rs1569135 CALCRL 2 A 0.53 5.60E-10 0.021 (20)
rs10804591 PLXND1 3 A 0.79 6.60E-09 0.025 (20)
rs17451107 LEKR1 3 T 0.61 1.10E-12 0.026 (20)
rs9991328 FAM13A 4 T 0.49 4.50E-08 0.019 (20)
rs303084 SPATA5-FGF2 4 A 0.8 3.90E-08 0.023 (20)
rs9687846 MAP3K1 5 A 0.19 7.10E-08 0.024 (20)
rs6556301 FGFR4 5 T 0.36 2.60E-08 0.022 (20)
rs7759742 BTNL2 6 A 0.51 4.40E-11 0.023 (20)
rs7801581 HOXA11 7 T 0.24 3.70E-10 0.027 (20)
rs7830933 NKX2-6 8 A 0.77 7.40E-08 0.022 (20)
rs12679556 MSC 8 G 0.25 2.10E-11 0.027 (20)
rs10991437 ABCA1 9 A 0.11 1.00E-08 0.031 (20)
rs11231693 MACROD1-VEGFB 11 A 0.06 4.50E-08 0.041 (20)
rs4765219 CCFC92 12 C 0.67 1.60E-15 0.028 (20)
rs8042543 KLF13 15 C 0.78 1.20E-09 0.026 (20)
rs8030605 RFX7 15 A 0.14 8.80E-09 0.03 (20)
rs1440372 SMAD6 15 C 0.71 1.10E-10 0.024 (20)
rs4646404 PEMT 17 G 0.67 1.40E-11 0.027 (20)
rs12608504 BCL2 18 A 0.36 8.80E-10 0.022 (20)
rs4081724 CEBPA 19 G 0.85 7.40E-12 0.035 (20)
rs979012 BMP2 20 T 0.34 3.30E-14 0.027 (20)
rs224333 GDF5 20 G 0.62 2.60E-08 0.02 (20)
rs6090583 EYA2 20 A 0.48 6.20E-11 0.022 (20)
rs2645294 TBX15-WARS2 1 T 0.58 1.70E-19 0.031 (20)
rs714515 DNM3-PIGC 1 G 0.43 4.40E-15 0.027 (20)
rs2820443 LYPLAL1 1 T 0.72 5.30E-21 0.035 (20)
rs10195252 GRB14-COBLL1 2 T 0.59 5.90E-15 0.027 (20)
rs17819328 PPARG 3 G 0.43 2.40E-09 0.021 (20)
rs2276824 PBRM1 3 C 0.43 3.20E-11 0.024 (20)
rs2371767 ADAMTS9 3 G 0.72 1.60E-20 0.036 (20)
rs7705502 CPEB4 5 A 0.33 4.70E-14 0.027 (20)
rs1294410 LY86 6 C 0.63 2.00E-18 0.031 (20)
rs1358980 VEGFA 6 T 0.47 3.10E-27 0.039 (20)
rs1936805 RSPO3 6 T 0.51 3.60E-35 0.043 (20)
rs10245353 NFE2L3 7 A 0.2 8.40E-16 0.035 (20)
rs10842707 ITPR2-SSPN 12 T 0.23 4.40E-16 0.032 (20)
rs1443512 HOXC13 12 A 0.24 6.90E-13 0.028 (20)
rs2294239 ZZNRF3 22 A 0.59 7.20E-13 0.025 (20)
Schizophrenia
Top SNP Risk allele RAF p-value OR References
rs4648845 C 0.527 8.70E-10 1.072 (21)
rs1498232 C 0.296 2.86E-09 1.069 (21)
rs11210892 A 0.323 3.39E-10 1.071 (21)
rs12129573 C 0.358 2.03E-12 1.078 (21)
rs76869799 C 0.036 2.64E-08 1.182 (21)
rs140505938 T 0.836 4.49E-10 1.094 (21)
rs6670165 C 0.184 4.45E-08 1.075 (21)
rs7523273 G 0.685 4.47E-08 1.063 (21)
rs10803138 A 0.762 2.03E-08 1.072 (21)
rs11682175 T 0.458 1.47E-11 1.072 (21)
rs3768644 A 0.899 7.39E-09 1.106 (21)
rs2909457 A 0.407 4.62E-08 1.059 (21)
rs11693094 T 0.542 1.53E-12 1.076 (21)
rs59979824 A 0.663 8.41E-09 1.067 (21)
rs6434928 A 0.357 2.06E-11 1.076 (21)
rs6704641 G 0.805 8.33E-09 1.081 (21)
rs11685299 A 0.674 1.12E-08 1.065 (21)
rs6704768 A 0.448 2.32E-12 1.075 (21)
rs17194490 G 0.156 2.69E-11 1.101 (21)
rs4330281 T 0.52 4.64E-09 1.064 (21)
rs75968099 C 0.324 1.05E-13 1.085 (21)
rs2535627 C 0.529 4.26E-11 1.071 (21)
rs832187 T 0.385 1.43E-08 1.063 (21)
rs7432375 A 0.551 7.26E-11 1.072 (21)
rs9841616 A 0.833 2.35E-08 1.081 (21)
rs215411 T 0.314 3.06E-08 1.064 (21)
rs35518360 A 0.078 7.98E-15 1.167 (21)
rs10520163 C 0.47 1.47E-09 1.065 (21)
rs1106568 A 0.239 9.47E-09 1.071 (21)
rs1501357 T 0.198 5.05E-09 1.08 (21)
rs4391122 A 0.468 1.10E-14 1.085 (21)
rs16867576 G 0.883 4.61E-09 1.101 (21)
rs4388249 C 0.213 3.05E-08 1.076 (21)
rs10043984 C 0.252 1.09E-08 1.069 (21)
rs79212538 G 0.046 7.00E-09 1.155 (21)
rs11740474 A 0.379 3.15E-08 1.062 (21)
rs115329265 G 0.85 3.48E-31 1.205 (21)
rs1339227 T 0.632 2.69E-08 1.062 (21)
rs117074560 T 0.9524 1.64E-09 1.178 (21)
rs12704290 A 0.877 3.33E-10 1.106 (21)
rs6466055 C 0.332 1.13E-09 1.068 (21)
rs211829 C 0.628 3.71E-08 1.061 (21)
rs13240464 C 0.647 3.03E-13 1.083 (21)
rs7801375 A 0.848 4.42E-08 1.082 (21)
rs3735025 C 0.642 3.28E-09 1.066 (21)
rs10503253 C 0.219 1.06E-08 1.073 (21)
rs73229090 A 0.884 2.10E-08 1.101 (21)
rs6984242 A 0.4 5.97E-09 1.063 (21)
rs7819570 G 0.174 1.22E-08 1.079 (21)
rs36068923 A 0.197 2.61E-11 1.088 (21)
rs4129585 C 0.424 1.74E-15 1.087 (21)
rs11139497 T 0.337 3.61E-09 1.069 (21)
rs7893279 G 0.889 1.97E-12 1.125 (21)
rs7907645 G 0.888 1.27E-11 1.143 (21)
rs11027857 G 0.499 2.55E-09 1.064 (21)
rs9420 G 0.311 2.24E-09 1.068 (21)
rs12421382 T 0.666 3.70E-08 1.063 (21)
rs2514218 T 0.686 2.75E-11 1.079 (21)
rs77502336 G 0.322 7.54E-09 1.066 (21)
rs55661361 A 0.665 2.80E-12 1.08 (21)
rs10791097 G 0.46 1.09E-12 1.076 (21)
rs75059851 G 0.797 3.87E-11 1.091 (21)
rs2007044 A 0.376 3.22E-18 1.096 (21)
rs679087 A 0.663 3.91E-08 1.063 (21)
rs324017 A 0.691 2.13E-08 1.066 (21)
rs4240748 C 0.634 4.59E-08 1.06 (21)
rs10860964 C 0.646 4.84E-08 1.06 (21)
rs4766428 C 0.474 1.40E-09 1.068 (21)
rs2851447 C 0.259 1.86E-14 1.093 (21)
rs2068012 T 0.229 1.41E-08 1.072 (21)
rs2332700 G 0.249 4.86E-09 1.073 (21)
rs2693698 A 0.582 4.80E-09 1.065 (21)
rs12887734 G 0.287 1.36E-13 1.088 (21)
rs56205728 G 0.274 4.18E-09 1.074 (21)
rs12903146 G 0.52 3.38E-10 1.067 (21)
rs12148337 C 0.465 1.79E-08 1.06 (21)
rs8042374 G 0.725 2.44E-13 1.093 (21)
rs950169 T 0.743 1.62E-11 1.083 (21)
rs4702 A 0.438 8.30E-14 1.085 (21)
rs9922678 G 0.281 1.28E-08 1.067 (21)
rs7405404 C 0.223 1.01E-09 1.077 (21)
rs12691307 G 0.51 4.55E-11 1.073 (21)
rs12325245 A 0.141 1.87E-08 1.087 (21)
rs8044995 G 0.162 1.51E-08 1.081 (21)
rs4523957 G 0.627 2.86E-10 1.071 (21)
rs8082590 A 0.386 1.77E-08 1.065 (21)
rs78322266 G 0.0292 1.32E-08 1.188 (21)
rs72934570 T 0.9203 1.97E-11 1.145 (21)
rs2905426 T 0.372 3.63E-10 1.071 (21)
rs2053079 A 0.231 4.49E-09 1.074 (21)
rs56873913 G 0.766 4.69E-08 1.071 (21)
rs6065094 A 0.678 1.46E-11 1.078 (21)
rs7267348 T 0.246 4.56E-08 1.067 (21)
rs9607782 T 0.232 2.07E-11 1.087 (21)
rs1023500 C 0.81 3.43E-08 1.076 (21)
rs12845396 A 0.247 2.21E-08 1.056 (21)
rs1378559 C 0.831 1.61E-12 1.09 (21)
rs5937157 T 0.241 1.98E-10 1.066 (21)
* Associated with diastolic blood pressure
Supplementary Table 2. Number of single nucleotide polymorphisms included in the different GRSs. Correlation coefficient between the GRSs in the MIGen study.
GRSs including all the genetic variants of interest
SNPs Ntotal
GRSIHD GRSTG GRSBP GRSHDL GRSLDL GRST2D GRSBMI GRSWaist GRSSCHIZ
GRSIHD 60 - 0.079 -0.010 -0.034 0.182 -0.002 0.002 -0.007 0.023 GRSTG 44 - -0.014 -0.391 0.170 0.043 0.014 0.055 -0.004 GRSBP 28 - 0.010 0.011 -0.013 -0.017 0.035 -0.014 GRSHDL 74 - -0.129 -0.049 -0.024 -0.045 0.006 GRSLDL 58 - 0.008 -0.025 0.000 -0.015 GRST2D 23 - 0.001 -0.046 -0.007 GRSBMI 81 - 0.029 -0.015 GRSWaist 41 - -0.024 GRSSCHIZ 98 -
GRSs excluding those genetic variants associated with any other trait different to that of interest
SNPs
Ntotal GRSIHD GRSTG GRSBP GRSHDL GRSLDL GRST2D GRSBMI GRSWaist GRSSCHIZ
GRSIHD 19 - 0.009 -0.022 -0.007 -0.062 -0.025 -0.050 0.008 .0.002 GRSTG 17 - 0.002 -0.142 0.379 0.002 -0.003 0.123 0.007 GRSBP 11 - 0.049 -0.011 -0.006 -0.074 0.034 0.011 GRSHDL 34 - -0.008 -0.043 -0.032 -0.083 0.032 GRSLDL 13 - 0.007 0.014 -0.008 -0.003 GRST2D 11 - -0.017 -0.022 -0.042 GRSBMI 32 - -0.010 -0.018 GRSWaist 16 - 0.013 GRSSCHIZ 20 -
GRS: genetic risk scores; SNPs: Single nucleotide polymorphisms; IHD: Ischemic heart disease; TG: Triglycerides; BP: Blood pressure; HDL: High density lipoprotein; LDL: Low density lipoprotein; T2D: Type 2 diabetes; BMI: Body Mass Index; Schiz: Schizophrenia.
Supplementary Table 3. Association Between the Weighted Genetic Risk Scores and Their
Corresponding Risk Factors in the Cohorts Included in this Analysis.
GRSs including all the genetic variants of interest
FHS* FINRISK97 FINRISK02 EGCUT*
GRSLDL-*
LDL (mg/dL)
7.16 (0.61)
p-value:2.3x10-31
8.93 (0.44)
p-value: 3.1x10-90
14.60 (0.95)
p-value: 6.3x10-51
8.92 (0-59)
p-value: 8.4x10-51
GRSHDL-*
HDL (mg/dL)
-3.58 (0.24)
p-value: 4.9x10-47
-3.48 (0.16)
p-value: 9.5x10-101
-5.54 (0.35)
p-value: 1.1x10-54
-3.56 (0.24)
p-value: 8.4x10-49
GRSTG-*
TG (mg/dL)
23.80 (1.82)
p-value: 5.2x10-38
18.62 (1.13)
p-value: 2.0x10-59
22.11 (1.83)
p-value:1.1x10-32
18.14 (1.43)
p-value: 3.3x10-36
GRSBP-*
SBP (mmHg)
1.43 (0.31)
p-value: 3.7x10-6
0.86 (0.23)
p-value: 1.9x10-4
0.88 (0.39)
p-value: 0.023
0.53 (0.20)
p-value: 6.7x10-3
GRSBP-*
DBP (mmHg)
0.54 (0.18)
p-value: 3.4x10-3
0.30 (0.14)
p-value: 0.031
0.39 (0.23)
p-value: 0.088
0.41 (0.13)
p-value: 1.5x10-3
GRST2D-*
T2D (yes)
0.34 (0.08)
p-value: 1.6x10-5
0.16 (0.06)
p-value: 3.2x10-3
0.14 (0.07)
p-value: 0.027
0.13 (0.07)
p-value:0.050
GRSBMI-*
BMI (kg/m2)
0.60 (0.09)
p-value: 1.8x10-11
0.60 (0.06)
p-value: 1.2x10-25
0.72 (0.10)
p-value: 4.7x10-12
0.71 (0.06)
p-value:5.0x10-30
GRSs excluding those genetic variants associated with any other trait different to that of interest
FHS* FINRISK97 FINRISK02 EGCUT*
GRSLDL-*
LDL (mg/dL)
0.94 (0.62)
p-value:0.129
2.35 (0.45)
p-value: 1.9x10-7
4.38 (0.99)
p-value: 1.1x10-5
2.02 (0.60)
p-value: 8.0x10-4
GRSHDL-*
HDL (mg/dL)
-1.59 (0.25)
p-value: 2.9x10-10
-1.58 (0.17)
p-value: 2.4x10-21
-3.03 (0.35)
p-value: 6.6x10-17
-1.51 (0.24)
p-value: 6.3x10-10
GRSTG-*
TG (mg/dL)
4.29 (1.87)
p-value: 0.022
8.32 (1.16)
p-value: 8.1x10-13
10.99 (1.87)
p-value:5.0x10-9
7.99 (1.43)
p-value: 2.4x10-8
GRSBP-*
SBP (mmHg)
1.28 (0.30)
p-value: 3.1x10-5
0.35 (0.23)
p-value: 0.132
-0.02 (0.39)
p-value: 0.954
0.00 (0.20)
p-value: 1.000
GRSBP-*
DBP (mmHg)
0.41 (0.18)
p-value: 0.023
0.15 (0.14)
p-value: 0.286
-0.18 (0.23)
p-value: 0.429
0.06 (0.13)
p-value: 0.662
GRST2D-*
T2D (yes)
0.22 (0.08)
p-value: 6.4x10-3
0.09 (0.06)
p-value: 0.103
0.16 (0.07)
p-value: 0.020
0.00 (0.07)
p-value:0.956
GRSBMI-*
BMI (kg/m2)
0.26 (0.09)
p-value: 3.6x10-3
0.35 (0.06)
p-value: 1.0x10-9
0.51 (0.10)
p-value: 1.0x10-6
0.50 (0.06)
p-value:1.3x10-15
* FHS: Framingham Heart Study; EGCUT: Estonian Biobank; GRS: Genetic risk score; LDL: Low-density lipoprotein; HDL: High-density lipoprotein; TG: Triglycerides; BP: Blood pressure; SBP: Systolic blood pressure; DBP: Diastolic blood pressure; T2D: Type 2 diabetes; BMI: Body Mass Index.
Supplementary Table 4. Association Between the Different Weighted Genetic Risk Scores for Cardiovascular Risk Factors and Ischemic Heart
Disease and the Prevalence/Incidence of Hard Ischemic Heart Disease Events (Myocardial Infarction or Ischemic Heart Disease Death).
MIGen* FHS* FINRISK1997 FINRISK2002 EGCUT* Meta-analysis p-het‡
GRS*
IHD*
OR-RR (95%CI)†
p-value
1.51 (1.42, 1.60)
8.1x10-39
1.37 (1.18, 1.59)
4.8x10-05
1.25 (1.14, 1.38)
2.4x10-06
1.26 (1.09, 1.46)
1.4x10-03
1.18 (1.04, 1.34)
9.5x10-03
1.37 (1.32, 1.43)
9.4x10-47
6.8
x10-4
GRS
TG*
OR-RR (95%CI)
p-value
1.16 (1.09, 1.23)
7.2x10-07
0.90 (0.77, 1.05)
0.164
1.02 (0.93, 1.11)
0.735
1.05 (0.91, 1.22)
0.492
1.07 (0.95, 1.20)
0.297
1.09 (1.04, 1.13)
1.1x10-4
0.010
GRS
BP*
OR-RR (95%CI)
p-value
1.03 (0.98, 1.10)
0.260
1.07 (0.92, 1.25)
0.371
1.10 (1.00, 1.21)
0.057
0.98 (0.85, 1.14)
0.833
0.96 (0.84, 1.08)
0.492
1.04 (0.99, 1.08)
0.107
0.460
GRS
HDL*
OR-RR (95%CI)
p-value
1.13 (1.07, 1.20)
3.2x10-05
0.85 (0.73, 0.99)
0.034
1.08 (0.98, 1.18)
0.103
1.00 (0.87, 1.16)
0.983
1.05 (0.93, 1.19)
0.396
1.08 (1.03, 1.12)
6.2x10-4
0.009
GRS
LDL*
OR-RR (95%CI)
p-value
1.21 (1.14, 1.29)
1.2x10-10
1.04 (0.89, 1.21)
0.660
1.05 (0.96, 1.15)
0.274
0.99 (0.86, 1.15)
0.945
0.99 (0.87, 1.12)
0.844
1.12 (1.07, 1.16)
3.1x10-7
0.002
GRS
T2D*
OR-RR (95%CI)
p-value
1.03 (0.97, 1.10)
0.266
1.03 (0.89, 1.19)
0.714
0.92 (0.84, 1.01)
0.083
1.10 (0.95, 1.27)
0.206
1.00 (0.88, 1.12)
0.947
1.01 (0.97, 1.05)
0.666
0.215
GRS
BMI*
OR-RR (95%CI)
p-value
1.13 (1.06, 1.19)
7.7x10-05
1.11 (0.95, 1.29)
0.179
1.05 (0.96, 1.15)
0.330
0.97 (0.84, 1.12)
0.709
1.03 (0.91, 1.16)
0.619
1.08 (1.04, 1.13)
2.3x10-4
0.271
GRS
WAIST
OR-RR (95%CI)
p-value
1.07 (1.00, 1.13)
0.035
0.99 (0.85, 1.19)
0.714
1.09 (0.99, 1.19)
0.072
1.00 (0.87, 1.16)
0.953
1.08 (0.96, 1.22)
0.196
1.06 (1.02, 1.11)
0.006
0.761
GRS
SCHIZ
OR-RR (95%CI)
p-value
1.01 (0.95, 1.07)
0.683
0.96 (0.82, 1.11)
0.571
0.95 (0.86, 1.04)
0.281
0.91 (0.78, 1.05)
0.188
0.98 (0.86, 1.11)
0.726
0.98 (0.94, 1.02)
0.395
0.594
*MIGen: Myocardial Infarction Genetics Consortium; FHS: Framingham Heart Study; EGCUT: Estonian Biobank; GRS: Genetic risk score; IHD: Ischemic heart disease; TG: Triglycerides; BP: Blood pressure; HDL: High-density lipoprotein; LDL: Low-density lipoprotein; T2D: Type 2 diabetes; BMI: Body Mass Index; Schiz: Schizophrenia. † OR-RR (95%CI): Odds ratio/Relative risks (and their 95% confidence intervals) corresponding to the increase in ischemic heart disease risk per standard deviation increase of the genetic risk score, and the P-values of the different reported associations. The OR was estimated in the MIGEN study and the RR in the rest of the studies.
‡ p-het: p-value for heterogeneity between studies.
Supplementary Table 5. Association Between the Weighted Genetic Risk Scores for Cardiovascular Risk Factors and Ischemic Heart Disease and
the Incidence of All Ischemic Heart Disease Events (myocardial infarction or ischemic heart disease death or angina or revascularization).
MIGen* FHS* FINRISK1997 FINRISK2002 EGCUT* Meta-analysis p-het‡
GRS*
IHD*
RR (95%CI)†
p-value
--- 1.32 (1.17, 1.49)
9.4x10-06
1.23 (1.11, 1.37)
9.6x10-05
1.29 (1.10, 1.52)
1.8x10-03
1.08 (0.99, 1.17)
0.066
1.19 (1.13, 1.25)
7.9x10-03
0.026
GRS
TG*
RR (95%CI)
p-value
--- 0.98 (0.87, 1.11)
0.773
0.98 (0.88, 1.08)
0.663
1.11 (0.94, 1.31)
0.201
1.06 (0.98, 1.15)
0.157
1.03 (0.97, 1.08)
0.324
0.403
GRS
BP*
RR (95%CI)
p-value
--- 1.03 (0.91, 1.17)
0.593
1.13 (1.02, 1.26)
0.021
0.88 (0.74, 1.04)
0.121
1.04 (0.95, 1.12)
0.394
1.04 (0.99, 1.10)
0.138
0.087
GRS
HDL*
RR (95%CI)
p-value
--- 0.92 (0.81, 1.04)
0.189
1.07 (0.97, 1.19)
0.181
1.06 (0.90, 1.26)
0.463
0.96 (0.89, 1.04)
0.353
0.99 (0.94, 1.05)
0.819
0.182
GRS
LDL*
RR (95%CI)
p-value
--- 1.05 (0.93, 1.19)
0.431
1.05 (0.95, 1.16)
0.331
1.01 (0.86, 1.19)
0.912
0.98 (0.90, 1.07)
0.680
1.02 (0.96, 1.07)
0.536
0.715
GRS
T2D*
RR (95%CI)
p-value
--- 1.05 (0.93, 1.19)
0.435
0.91 (0.82, 1.00)
0.057
1.19 (1.01, 1.41)
0.034
1.01 (0.93, 1.09)
0.850
1.00 (0.95, 1.06)
0.889
0.033
GRS
BMI*
RR (95%CI)
p-value
--- 1.13 (1.00, 1.28)
0.051
1.08 (0.98, 1.20)
0.127
1.00 (0.85, 1.17)
0.969
0.97 (0.90, 1.05)
0.475
1.03 (0.98, 1.09)
0.246
0.150
GRS
WAIST
RR (95%CI)
p-value
--- 1.06 (0.93, 1.19)
0.394
1.09 (0.99, 1.21)
0.087
1.01 (0.86, 1.19)
0.890
1.10 (1.01, 1.19)
0.024
1.08 (1.02, 1.14)
0.005
0.801
GRS
SCHIZ*
RR (95%CI)
p-value
--- 0.99 (0.87, 1.12)
0.828
0.96 (0.86, 1.06)
0.414
0.91 (0.77, 1.07)
0.262
0.99 (0.91, 1.07)
0.773
0.97 (0.92, 1.03)
0.283
0.824
*MIGen: Myocardial infarction Genetics Consortium; FHS: Framingham Heart Study; EGCUT: Estonian Biobank; GRS: Genetic risk score; IHD: Ischemic heart disease; TG: Triglycerides; BP: Blood pressure; HDL: High-density lipoprotein; LDL: Low-density lipoprotein; T2D: Type 2 diabetes; BMI: Body Mass Index; Schiz: Schizophrenia. † OR-RR (95%CI): Odds ratio/Relative risks (and their 95% confidence intervals) corresponding to the increase in ischemic heart disease risk per standard deviation increase of the genetic risk score, and the P-values of the different reported associations. The OR was estimated in the MIGEN study and the RR in the rest of the studies.
‡ p-het: p-value for heterogeneity between studies.
Supplementary Table 6. Association Between the Non-pleiotropic Weighted Genetic Risk Scores and the
Prevalence/Incidence of Ischemic Heart Disease Events in the Meta-analyses.
Hard coronary events All coronary events
Non-pleiotropic GRS* IHD* RR (95%CI)†
p-value
1.28 (1.23, 1.34)
6.8x10-30
1.12 (1.06,1.18)
2.4x10-5
Non-pleiotropic GRS TG* RR (95%CI)
p-value
1.02 (0.98, 1.06)
0.391
1.00 (0.94,1.06)
0.960
Non-pleiotropic GRS BP* RR (95%CI)
p-value
1.0 (0.96, 1.04)
0.981
1.03 (0.97, 1.08)
0.339
Non-pleiotropic GRS HDL* RR (95%CI)
p-value
1.04 (1.00, 1.08)
0.074
1.00 (0.95, 1.06)
0.995
Non-pleiotropic GRS LDL* RR (95%CI)
p-value
1.0 (0.96, 1.04)
0.919
0.99 (0.94, 1.04)
0.696
Non-pleiotropic GRS T2D* RR (95%CI)
p-value
1.02 (0.98,1.06)
0.335
1.00 (0.95, 1.06)
0.875
Non-pleiotropic GRS BMI* RR (95%CI)
p-value
1.03 (0.98, 1.07)
0.230
1.02 (0.96, 1.07)
0.576
Non-pleiotropic GRS Waist RR (95%CI)
p-value
1.04 (0.99, 1.08)
0.132
1.04 (0.98, 1.09)
0.189
Non-pleiotropic GRS SCHIZ* RR (95%CI)
p-value
0.96 (0.92, 1.00)
0.083
0.96 (0.91, 1.01)
0.129
*GRS: Genetic risk score; IHD: Ischemic heart disease; TG: Triglycerides; BP: Blood pressure; HDL: High-density lipoprotein; LDL: Low-density lipoprotein; T2D: Type 2 diabetes; BMI: Body Mass Index; Schiz: Schizophrenia. † 95%CI: 95% confidence intervals.
Supplementary Table 7. Results of the Meta-analyses for the Multiplicative Interaction Between the Different Genetic Risk Scores of Interest on
Ischemic Heart Disease Risk (Hard Events). The Regression Coefficient of the Interaction Terms (Standard Error) is Shown in the Upper-right Part
of the Diagonal, and the p-Value of the Multiplicative Interaction and the p-Value for the Heterogeneity Between the Analyzed Studies is Shown in
the Lower-left Part of the Diagonal.
GRS-IHD* GRS-TG* GRS-BP* GRS-HDL* GRS-LDL* GRS-T2D* GRS-BMI* GRS-Waist GRS-SCHIZ
GRS-IHD NA -0.003
(0.022)
-0.008
(0.022)
-0.003
(0.022)
-0.005
(0.022)
0.022
(0.022)
0.012
(0.022)
0.015
(0.022)
-0.021
(0.022)
GRS-TG p-value=0.898
p-het†=0.097 NA
-0.028
(0.022)
-0.017
(0.020)
-0.047
(0.021)
0.029
(0.021)
0.035
(0.021)
-0.020
(0.022)
-0.018
(0.022)
GRS-BP p-value=0.712
p-het=0.411
p-value=0.204
p-het=0.974 NA
-0.021
(0.022)
-0.040
(0.022)
-0.019
(0.022)
-0.013
(0.022)
-0.041
(0.022)
-0.015
(0.022)
GRS-HDL p-value=0.883
p-het=0.308
p-value=0.406
p-het=0.776
p-value=0.334
p-het=0.695 NA
-0.015
(0.022)
-0.002
(0.021)
0.020
(0.021)
-0.013
(0.021)
-0.024
(0.022)
GRS-LDL p-value=0.816
p-het=0.802
p-value=0.027
p-het=0.011
p-value=0.066
p-het=0.207
p-value=0.473
p-het=0.228 NA
-0.024
(0.022)
-0.025
(0.021)
-0.009
(0.022)
-0.003
(0.021)
GRS-T2D p-value=0.305
p-het=0.322
p-value=0.179
p-het=0.065
p-value=0.392
p-het=0.653
p-value=0.909
p-het=0.138
p-value=0.259
p-het=0.103 NA
-0.005
(0.021)
-0.003
(0.022)
0.024
(0.022)
GRS-BMI p-value=0.589
p-het=0.385
p-value=0.102
p-het=0.179
p-value=0.540
p-het=0.583
p-value=0.340
p-het=0.788
p-value=0.242
p-het=0.781
p-value=0.828
p-het=0.577 NA
-0.007
(0.021)
-0.009
(0.022)
GRS-Waist p-value=0.483
p-het=0.430
p-value=0.368
p-het=0.379
p-value=0.059
p-het=0.903
p-value=0.537
p-het=0.023
p-value=0.663
p-het=0.368
p-value=0.534
p-het=0.703
p-value=0.730
p-het=0.913
NA -0.034
(0.021)
GRS-SCHIZ p-value=0.355
p-het=0.885
p-value=0.408
p-het=0.067
p-value=0.501
p-het=0.242
p-value=0.265
p-het=0.435
p-value=0.901
p-het=0.506
p-value=0.283
p-het=0.186
p-value=0.665
p-het=0.399
p-value=0.111
p-het=0.749
NA
*GRS: Genetic risk score; IHD: Ischemic Heart Disease; TG: Triglycerides; BP: Blood pressure; HDL: High-density lipoprotein; LDL: Low-density lipoprotein; T2D: Type 2 diabetes; BMI: Body Mass Index; Schiz: Schizophrenia. †P-het: p value for the heterogeneity test.
Supplementary Table 8. Results of the Meta-analyses for the Multiplicative Interaction Between the Different Genetic Risk Scores of Interest on
Ischemic Heart Disease Risk (All Events). The Regression Coefficient of the Interaction Terms (Standard Error) is Shown in the Upper-right Part of
the Diagonal, and the p-Value of the Multiplicative Interaction and the p-Value for the Heterogeneity Between the Analyzed Studies is Shown in the
Lower-left Part of the Diagonal.
GRS-IHD* GRS-TG* GRS-BP* GRS-HDL* GRS-LDL* GRS-T2D* GRS-BMI* GRS-Waist
GRS-
SCHIZ
GRS-IHD NA -0.024
(0.028)
0.022
(0.028)
0.018
(0.028)
0.003
(0.027)
0.070
(0.027)
0.048
(0.027)
0.036
(0.028)
0.012
(0.029)
GRS-TG p-value=0.395
p-het†=0.517 NA
-0.046
(0.028)
-0.015
(0.025)
-0.096
(0.028)
0.034
(0.027)
0.014
(0.027)
0.026
(0.027)
-0.040
(0.028)
GRS-BP p-value=0.423
p-het=0.881
p-value=0.099
p-het=0.629 NA
0.001
(0.027)
-0.023
(0.028)
0.008
(0.027)
0.009
(0.027)
-0.067
(0.027)
0.018
(0.028)
GRS-HDL p-value=0.513
p-het=0.666
p-value=0.562
p-het=0.909
p-value=0.973
p-het=0.952 NA
-0.046
(0.027)
0.001
(0.028)
0.008
(0.027)
0.059
(0.027)
-0.046
(0.028)
GRS-LDL p-value=0.922
p-het=0.478
p-value=5x10-4
p-het=0.252
p-value=0.397
p-het=0.978
p-value=0.091
p-het=0.541 NA
-0.018
(0.028)
-0.016
(0.027)
0.039
(0.026)
0.002
(0.027)
GRS-T2D p-value=0.009
p-het=0.495
p-value=0.211
p-het=0.063
p-value=0.757
p-het=0.880
p-value=0.972
p-het=0.180
p-value=0.513
p-het=0.235 NA
-0.023
(0.027)
0.015
(0.027)
0.047
(0.028)
GRS-BMI p-value=0.078
p-het=0.532
p-value=0.601
p-het=0.217
p-value=0.748
p-het=0.509
p-value=0.779
p-het=0.662
p-value=0.565
p-het=0.693
p-value=0.393
p-het=0.702 NA
-0.013
(0.026)
0.026
(0.028)
GRS-Waist p-value=0.191
p-het=0.061
p-value=0.343
p-het=0.206
p-value=0.014
p-het=0.634
p-value=0.025
p-het=0.050
p-value=0.131
p-het=0.434
p-value=0.576
p-het=0.101
p-value=0.609
p-het=0.260
NA -0.020
(0.027)
GRS-SCHIZ p-value=0.678
p-het=0.942
p-value=0.144
p-het=0.026
p-value=0.515
p-het=0.134
p-value=0.097
p-het=0.504
p-value=0.944
p-het=0.676
p-value=0.092
p-het=0.576
p-value=0.343
p-het=0.727
p-value=0.455
p-het=0.442
NA
*GRS: Genetic risk score; IHD: Ischemic Heart Disease; TG: Triglycerides; BP: Blood pressure; HDL: High-density lipoprotein; LDL: Low-density lipoprotein; T2D: Type 2 diabetes; BMI: Body Mass Index; Schiz: Schizophrenia. †P-het: p value for the heterogeneity test.
Supplementary Table 9. Results of the sensitivity analyses undertaken to assess whether an
individual study influenced the pooled results of the multiplicative GRSLDL·GRSTG and
GRSLDL·GRSIHD interaction terms, excluding one study at a time and calculating the multiplicative
interaction metrics for the remaining studies considering hard and all coronary events.
Hard coronary events
Regression coefficient
(Standard Error)
All coronary events
Regression coefficient
(Standard Error)
GRSLDL·GRSTG
GRS global analysis -0.047 (0.021) -0.096 (0.028)
Global–Excluding MIGEN -0.116 (0.031) ---
Global–Excluding Framingham -0.041 (0.022) -0.097 (0.030)
Global–Excluding FINRISK-1997 -0.018 (0.024) -0.071 (0.032)
Global–Excluding FINRISK-2002 -0.039 (0.022) -0.088 (0.029)
Global–Excluding EGCUT -0.049 (0.023) -0.142 (0.037)
Non-pleiotropic GRSIHD·GRSLDL
GRS global analysis 0.064 (0.022) 0.091 (0.028)
Global–Excluding MIGEN 0.079 (0.031) ---
Global–Excluding Framingham 0.058 (0.023) 0.084 (0.031)
Global–Excluding FINRISK-1997 0.073 (0.025) 0.104 (0.033)
Global–Excluding FINRISK-2002 0.072 (0.023) 0.102 (0.030)
Global–Excluding EGCUT 0.047 (0.024) 0.065 (0.037)
GRS: Genetic risk score; LDL: Low density lipoprotein; TG: Triglycerides; IHD: Ischemic Heart
Disease.
Supplementary Table 10. Results of the Meta-analyses for the Additive Interaction Between the Different Genetic Risk Scores of Interest on
Ischemic Heart Disease Risk (Hard Events). The Synergy Index (and the 95% Confidence Interval) is Shown in the Upper-right Part of the Diagonal;
the p-Value of the Additive Interaction and the p-Value for the Heterogeneity Between the Analyzed Studies is Shown in the Lower-left Part of the
Diagonal.
GRS-IHD* GRS-TG* GRS-BP* GRS-HDL* GRS-LDL* GRS-T2D* GRS-BMI* GRS-Waist GRS-SCHIZ
GRS-
IHD NA
1.063
(0.919, 1.207)
0.989
(0.842, 1.136)
0.975
(0.803, 1.147)
1.076
(0.933, 1.219)
1.038
(0.883, 1.193)
1.084
(0.938, 1.231)
1.122
(0.981, 1.264)
0.894
(0.736, 1.052)
GRS-
TG
p-value=0.392
p-het†=0.063 NA
0.865
(0.552, 1.177)
1.009
(0.499, 1.519)
1.137
(0.900, 1.374)
1.083
(0.760, 1.406)
1.158
(0.880, 1.437)
0.834
(0.523, 1.145)
1.015
(0.664, 1.367)
GRS-
BP
p-value=0.879
p-het=0.357
p-value=0.397
p-het=0.983 NA
0.802
(0.338, 1.265)
0.673
(0.385, 0.961)
0.786
(0.174, 1.397)
0.919
(0.589, 1.248)
0.755
(0.332, 1.179)
0.831
(0.060, 1.602)
GRS-
HDL
p-value=0.779
p-het=0.196
p-value=0.973
p-het=0.970
p-value=0.402
p-het=0.974 NA
0.793
(0.278, 1.308)
0.824
(0.455, 1.193)
0.839
(0.368, 1.310)
0.453
(-0.265, 1.170)
1.004
(0.642, 1.366)
GRS-
LDL
p-value=0.297
p-het=0.399
p-value=0.257
p-het=0.443
p-value=0.026
p-het=0.965
p-value=0.431
p-het=0.882 NA
0.903
(0.612, 1.194)
0.970
(0.755, 1.186)
0.912
(0.666, 1.157)
0.993
(0.683, 1.302)
GRS-
T2D
p-value=0.630
p-het=0.418
p-value=0.613
p-het=0.623
p-value=0.492
p-het=0.875
p-value=0.351
p-het=0.542
p-value=0.513
p-het=0.877 NA
0.957
(0.578, 1.335)
0.715
(0.057, 1.373)
0.831
(0.071, 1.591)
GRS-
BMI
p-value=0.256
p-het=0.669
p-value=0.266
p-het=0.905
p-value=0.628
p-het=0.888
p-value=0.503
p-het=0.990
p-value=0.788
p-het=0.920
p-value=0.822
p-het=0.902 NA
0.994
(0.690, 1.299)
0.940
(0.474, 1.407)
GRS-
Waist
p-value=0.091
p-het=0.489
p-value=0.294
p-het=0.728
p-value=0.258
p-het=0.946
p-value=0.135
p-het=0.990
p-value=0.480
p-het=0.987
p-value=0.396
p-het=0.620
p-value=0.972
p-het=0.975 NA
0.645
(0.032, 1.258)
GRS-
SCHIZ
p-value=0.189
p-het=0.959
p-value=0.932
p-het=0.837
p-value=0.667
p-het=0.942
p-value=0.982
p-het=0.787
p-value=0.962
p-het=0.999
p-value=0.663
p-het=0.808
p-value=0.802
p-het=0.944
p-value=0.257
p-het=0.956 NA
*GRS: Genetic risk score for the traits of interest; IHD: Ischemic Heart Disease; TG: Triglycerides; BP: Blood pressure; HDL: High-density lipoprotein; LDL: Low-density lipoprotein; T2D: Type 2 diabetes; BMI: Body Mass Index; Schiz: Schizophrenia. †P-het: p value for the heterogeneity between studies.
Supplementary Table 11. Results of the Meta-analyses for the Additive Interaction Between the Different Genetic Risk Scores of Interest on
Ischemic Heart Disease Risk (All Events). The Synergy Index (and the 95% Confidence Interval) is Shown in the Upper-right Part of the Diagonal;
the p-Value of the Additive Interaction and the p-Value for the Heterogeneity Between the Analyzed Studies is Shown in the Lower-left Part of the
Diagonal.
GRS-IHD* GRS-TG* GRS-BP* GRS-HDL* GRS-LDL* GRS-T2D* GRS-BMI* GRS-Waist GRS-SCHIZ
GRS-
IHD NA
0.864
(0.566, 1.162)
1.118
(0.832, 1.404)
0.995
(0.646, 1.344)
0.911
(0.596, 1.227)
1.231
(0.902, 1.561)
1.221
(0.928, 1.514)
1.171
(0.898, 1.445)
0.970
(0.610, 1.330)
GRS-
TG
p-value=0.370
p-het†=0.515 NA
0.584
(-0.041, 1.209)
0.781
(-0.020, 1.583)
0.123
(-1.093, 1.340)
0.822
(0.404, 1.241)
1.320
(0.436, 2.203)
1.238
(0.662, 1.814)
0.426
(-1.019, 1.864)
GRS-
BP
p-value=0.418
p-het=0.942
p-value=0.192
p-het=0.961 NA
1.149
(0.563, 1.734)
0.880
(0.382, 1.378)
1.007
(0.170, 1.844)
1.017
(0.582, 1.453)
0.666
(0.286, 1.046)
1.270
(0.680, 1.859)
GRS-
HDL
p-value=0.977
p-het=0.770
p-value=0.593
p-het=0.996
p-value=0.618
p-het=0.986 NA
0.671
(-0.098, 1.442)
0.705
(-0.111, 1.522)
1.122
(0.578, 1.666)
0.824
(0.163, 1.485)
0.741
(0.028, 1.453)
GRS-
LDL
p-value=0.582
p-het=0.526
p-value=0.158
p-het=0.953
p-value=0.636
p-het=0.997
p-value=0.403
p-het=0.674 NA
1.221
(0.475, 1.968)
0.952
(0.460, 1.444)
1.159
(0.487, 1.831)
0.897
(-0.045, 1.839)
GRS-
T2D
p-value=0.169
p-het=0.509
p-value=0.406
p-het=0.741
p-value=0.987
p-het=0.928
p-value=0.479
p-het=0.998
p-value=0.561
p-het=0.751 NA
1.000
(0.367, 1.633)
1.104
(0.475, 1.733)
0.841
(0.062, 1.621)
GRS-
BMI
p-value=0.139
p-het=0.609
p-value=0.478
p-het=0.902
p-value=0.939
p-het=0.929
p-value=0.660
p-het=0.980
p-value=0.848
p-het=0.937
p-value=1.000
p-het=0.991 NA
0.901
(0.438, 1.365)
0.835
(0.036, 1.634)
GRS-
Waist
p-value=0.219
p-het=0.129
p-value=0.418
p-het=0.836
p-value=0.085
p-het=0.773
p-value=0.602
p-het=0.542
p-value=0.642
p-het=0.604
p-value=0.746
p-het=0.395
p-value=0.676
p-het=0.485 NA
0.741
(0.028, 1.453)
GRS-
SCHIZ
p-value=0.871
p-het=0.968
p-value=0.432
p-het=0.845
p-value=0.370
p-het=0.974
p-value=0.475
p-het=0.975
p-value=0.830
p-het=0.994
p-value=0.689
p-het=0.789
p-value=0.686
p-het=0.984
p-value=0.475
p-het=0.975 NA
*GRS: Genetic risk score for the traits of interest; IHD: Ischemic Heart Disease; TG: Triglycerides; BP: Blood pressure; HDL: High-density lipoprotein; LDL: Low-density lipoprotein; T2D: Type 2 diabetes; BMI: Body Mass Index; Schiz: Schizophrenia. †P-het: p value for the heterogeneity between studies.
Supplementary Figure 1. Forest Plot of the Association Between the Different Weighted Genetic Risk Scores for Cardiovascular Risk Factors and Ischemic Heart Disease and the Incidence of All Ischemic Heart Disease Events (Myocardial Infarction, Ischemic Heart Disease Death, Angina and Revascularization) Across Studies and in the Meta-analysis.
REFERENCES
1. Kathiresan S, Voight BF, Purcell S, Musunuru K, Ardissino D, Mannucci, et al. Genome-wide association of early-onset myocardial infarction with single nucleotide polymorphisms and copy number variants. Nat Genet. 2009;41:334–341.
2. Kannel WB, Dawber TR, Kagan A, Revotskie N, Stokes J III. Factors of risk in the development of Coronary Heart Disease—Six-Year Follow-up Experience. The Framingham Study. Ann Intern Med. 1961;55:33–50.
3. Larson MG, Atwood LD, Benjamin EJ, Cupples LA, D'Agostino RB Sr, Fox CS, et al. Framingham Heart Study 100K project: genome-wide associations for cardiovascular disease outcomes. BMC Med Genet. 2007;8-Suppl 1:S5.
4. Leitsalu L, Haller T, Esko T, Tammesoo ML, Alavere H, Snieder H, et al. Cohort Profile: Estonian Biobank of the Estonian Genome Center, University of Tartu. Int J Epidemiol. 2014;44:1137-1147.
5. VanderWeele TJ, Knol MJ. A Tutorial on interaction. Epidemiol Method. 2014;3:33–72.
6. Rothman KJ, Lash TL, Greenland S. Modern Epidemiology. Boston: Little, Brown and Company; 1986.
7. Poole C. On the origin of risk relativism. Epidemiology 2010;21:3–9. 8. Schunkert H, König IR, Kathiresan S, Reilly MP, Assimes TL, Holm H, et al.
Large-scale association analysis identifies 13 new susceptibility loci for coronary artery disease. Nat Genet. 2011;43:333-338.
9. Coronary Artery Disease (C4D) Genetics Consortium. A genome-wide association study in Europeans and South Asians identifies five new loci for coronary artery disease. Nat Genet. 2011;43:339-344.
10. CARDIoGRAMplusC4D Consortium, Deloukas P, Kanoni S, Willenborg C, Farrall M, Assimes TL, et al. Large-scale association analysis identifies new risk loci for coronary artery disease. Nat Genet. 2013;45:25-33.
11. Nikpay M, Goel A, Won HH, Hall LM, Willenborg C, Kanoni S, et al. A comprehensive 1,000 Genomes-based genome-wide association meta-analysis of coronary artery disease. Nat Genet. 2015;47:1121-1130.
12. Teslovich TM, Musunuru K, Smith AV, Edmondson AC, Stylianou IM, Koseki M, et al. Biological, clinical and population relevance of 95 loci for blood lipids. Nature 2010;466:707–713.
13. Global Lipid Genetics Consortium. Discovery and refinement of loci associated with lipid levels. Nat Gen. 2013:45:1274-1283.
14. International Consortium for Blood Pressure Genome-Wide Association Studies, Ehret GB, Munroe PB, Rice KM, Bochud M, Johnson AD, et al. Genetic variants in novel pathways influence blood pressure and cardiovascular disease risk. Nature 2011;478:103–109.
15. Wain LV, Verwoert GC, O'Reilly PF, Shi G, Johnson T, Johnson AD, et al. Genome-wide association study identifies six new loci influencing pulse pressure and mean arterial pressure. Nat Genet. 2011;43:1005–1011.
16. Voight BF, Scott LJ, Steinthorsdottir V, Morris AP, Dina C, Welch RP, et al. Twelve type 2 diabetes susceptibility loci identified through large-scale association analysis. Nat Genet. 2010;42:579–589.
17. Mahajan A, Go MJ, Zhang W, Below JE, Gaulton KJ, Ferreira T et al. Genome-wide trans-ancestry meta-analysis provides insight into the genetic architecture of type 2 diabetes susceptibility. Nat Genet. 2014;46:234-244.
18. Saxena R, Voight BF, Lyssenko V, Burtt NP, de Bakker PI, Chen H, et al. Genome-wide association analysis identifies loci for type 2 diabetes and triglyceride levels. Science 2007;316:1331-1336.
19. Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH, Day FR, et al. Genetic studies of body mass index yield new insights for obesity biology. Nature 2015;518:197-206.
20. Shungin D, Winkler TW, Croteau-Chonka DC, Ferreira T, Locke AE, Mägi R, et al. New genetic loci link adipose and insulin biology to body fat distribution. Nature 2015;518:187-196.
21. Schizophrenia Working Group of the Psychiatric Genomics Consortium. Biological insights from 108 schizophrenia-associated genetic loci. Nature 2014;511:421-427.