Supplementary Online Content - JAMA...© 2016 American Medical Association. All rights reserved. Supplementary Online Content Lotta LA, Sharp SJ, Burgess S, et al. Association between

© 2016 American Medical Association. All rights reserved.

Supplementary Online Content

Lotta LA, Sharp SJ, Burgess S, et al. Association between low-density lipoprotein cholesterol–lowering genetic variants and risk of type 2 diabetes: a meta-analysis. JAMA. doi:10.1001/jama.2016.14568

The EPIC-InterAct Consortium

eTable 1. Participating studies

eTable 2. Genetic variants included in the main analysis

eTable 3. Sensitivity analyses at the NPC1L1 and PCSK9 loci

eTable 4. Correlation between genetic variants

eTable 5. Burden of rare alleles in exome sequencing studies

eFigure 1. Meta-analysis results

eFigure 2. Conditional analysis at the NPC1L1 locus

eFigure 3. Conditional analysis at the PCSK9 locus

eFigure 4. Associations of LDL-lowering alleles with continuous cardiometabolic traits

eFigure 5. Stratified associations of NPC1L1 variants

eFigure 6. Associations with continuous cardiometabolic traits

eReferences

This supplementary material has been provided by the authors to give readers additional information about their work.

Downloaded From: https://jamanetwork.com/ by a Non-Human Traffic (NHT) User on 05/28/2021


The EPIC‐InterAct Consortium: Claudia Langenberg, MD, PhD; Robert A. Scott, PhD; Stephen J. Sharp; Eva Ardanaz, MD PhD; Larraitz Arriola MD MSc; Beverley Balkau, PhD; Heiner Boeing, PhD; Panos Deloukas, PhD; Nita G Forouhi, FFPHM; Paul W Franks, PhD; Sara Grioni, BSc; Rudolf Kaaks, PhD; Timothy J Key, DPhil; Carmen Navarro, MD PhD MSc; Peter M Nilsson, PhD; Kim Overvad, PhD; Domenico Palli, MD; Salvatore Panico, MD; Jose‐Ramón Quirós, MD; Elio Riboli, MD MPH, ScM; Olov Rolandsson, MD PhD; Carlotta Sacerdote, MD, PhD; Elena C Salamanca‐Fernandez, MSc; Nadia Slimani, PhD; Annemieke MW Spijkerman; Anne Tjonneland, DrMedSci; Rosario Tumino, MD MSc, DLSHTM; Daphne L van der A, PhD; Yvonne T van der Schouw, PhD; Mark I. McCarthy, MD; Inês Barroso, PhD; Nicholas J. Wareham, MB BS, PhD.



eTable 1. Participating studies

Outcome Participating study Cases,

N

Non-cases (for case-control studies) or

participants (for continuous traits

studies), N

PubMed ID for cohort description

Website (URL)

Type 2 diabetes –

main analysis

InterAct-GWAS1* 4,187 4,254 21717116 http://www.inter-act.eu/ InterAct-CoreExome1* 5,121 7,269 21717116 http://www.inter-act.eu/

UK Biobank2 6,627 143,765 22463865 http://www.ukbiobank.ac.uk/ DIAGRAM3 34,840 114,981 22885922 http://diagram-consortium.org/

Type 2 diabetes –

exome sequencing

analysis

T2D-GENES Consortium, GoT2D Consortium,

DIAGRAM Consortium4 8,373 8,466 27398621 http://www.type2diabetesgenetics.org/home/portalHome

Coronary artery

disease

CARDIoGRAMplusC4D Consortium5 60,801 123,504 26343387 http://www.cardiogramplusc4d.org/

LDL cholesterol

Global Lipids Genetics Consortium6 - 188,577 24097068 http://csg.sph.umich.edu//abecasis/public/lipids2013/

Fasting plasma glucose

MAGIC Consortium7,8 - 133,010 22885924, 22581228 http://www.magicinvestigators.org/

Fasting insulin MAGIC Consortium7,8 - 108,557 22885924,

22581228 http://www.magicinvestigators.org/

Two hour glucose MAGIC Consortium7,8 - 42,854 22885924,

20081857 http://www.magicinvestigators.org/

Body mass index GIANT Consortium9 - 333,495 25673413 https://www.broadinstitute.org/collaboration/giant/index.php/GIANT_consortium

Waist-to-hip ratio GIANT Consortium10 - 224,047 25673412 https://www.broadinstitute.org/collaboration/giant/index.php/GIANT_consortium

Abbreviations: N, number of participants; LDL, low-density lipoprotein cholesterol. *In EPIC-Interact, genotyping was performed in two batches using the Illumina 660w quad and Illumina CoreExome genotyping arrays. Therefore, results of the main analysis are presented separately for individuals genotyped with the Illumina 660w quad array (InterAct-GWAS sub-study; 4,187 type 2 diabetes cases and 4,254 non-cases from the subcohort) and for individuals genotyped with the Illumina CoreExome array (InterAct-CoreExome sub-study; 5,121 type 2 diabetes cases and 7,269 non-cases from the subcohort).



eTable 2. Genetic variants included in the main analysis

Gene Polymorphism Genomic position

Effect allele

Effect allele frequency, mean

(range)

Genotyped or imputed (imputation quality score*),

Interact-GWAS


Interact-CoreExome


UK Biobank

NPC1L1 rs2073547 chr7:44582331 A 0.81 (0.81, 0.82) Imputed (0.995) Imputed (0.998) Genotyped

NPC1L1 rs217386 chr7:44600695 A 0.42 (0.41, 0.44) Imputed (0.991) Imputed (0.998) Genotyped

HMGCR rs12916 chr5:74656539 T 0.58 (0.57, 0.60) Imputed (0.994) Genotyped Genotyped

HMGCR rs5744707 chr5:74890618 A 0.90 (0.90, 0.91) Genotyped Imputed (0.993) Imputed (0.996)

HMGCR rs16872526 chr5:74675717 T 0.91 (0.90, 0.92) Imputed (0.999) Imputed (0.998) Imputed (0.997)

PCSK9 rs11591147 chr1:55505647 T 0.02 (0.01, 0.02) Imputed (0.877) Genotyped Genotyped

ABCG5/G8 rs4299376 chr2:44072576 T 0.69 (0.68, 0.70) Genotyped Genotyped Imputed (0.995)

LDLR rs6511720 chr19:11202306 T 0.11 (0.10, 0.12) Genotyped Genotyped Genotyped

In DIAGRAM, genetic variants were directly genotyped in the Metabochip subset of the DIAGRAM meta-analysis and either directly genotyped or imputed in the genome-wide association subset.3 Polymorphism names reported in the table are rsID entries from dbSNP release 147. Genomic coordinates represent chromosome and physical position of genetic variants according to the Human Reference Genome Build 37. Effect allele frequency averages and ranges are from EPIC-InterAct,1 UK Biobank2 and DIAGRAM.3 *imputation quality score reports the correlation between genotyped and imputed genotypes in the reference imputation set, with a value of 1 indicating perfect imputation.



eTable 3. Sensitivity analyses at the NPC1L1 and PCSK9 loci

Locus Model Reference

(PubMed ID) Polymorphisms

OR of type 2 diabetes

(95% CI)

p-value

NPC1L1 Two polymorphisms, adjusted for the GCK rs1799884 and rs2041547 lead

genetic variants This study rs2073547

rs217386 2.16

(1.51 – 3.11) 3 x

10-05

NPC1L1 Five polymorphisms Ference et al. (25770315)11

rs2073547 rs217386 rs7791240 rs2300414 rs10234070

2.20 (1.59 – 3.05)

2 x 10-06

PCSK9 Two polymorphisms This study rs11591147 rs471705 1.21 (1.04 – 1.41) 0.01

PCSK9 Nine polymorphisms This study

rs11591147 rs1998013 rs11206510 rs7523242 rs4927207 rs6662286 rs572512

rs1475701 rs7552841

1.16 (1.03 – 1.31) 0.02

Association with type 2 diabetes of LDL-cholesterol lowering genetic variants at the NPC1L1 and PCSK9 loci in sensitivity analyses. Odds ratios are per a genetically-predicted reduction in LDL cholesterol of 1 mmol/L. OR, odds ratio; CI, confidence interval.



eTable 4. Correlation between genetic variants Correlation between genetic variants included in Mendelian randomization models at the NPC1L1, HMGCR and PCSK9 loci. The correlation between variants was obtained from the SNAP software12 or from the 1000 Genomes Project browser (URL:http://browser.1000genomes.org/).

Genetic variant 1 Genetic variant 2 Correlation NPC1L1 locus

rs2073547 rs10234070 0.294 rs2073547 rs7791240 0.208 rs2073547 rs2300414 0.098 rs2073547 rs217386 0.083 rs217386 rs7791240 0.078 rs217386 rs10234070 0.06 rs217386 rs2300414 0.052 rs7791240 rs2300414 0.363 rs7791240 rs10234070 0.005 rs2300414 rs10234070 0.033

HMGCR locus rs12916 rs17238484 0.368 rs12916 rs5744707 0.239 rs12916 rs16872526 0.082

rs5744707 rs17238484 0.036 rs5744707 rs16872526 0.008

rs16872526 rs17238484 0.222 PCSK9 locus

rs11591147 rs471705 0.028 rs11591147 rs1998013 0.300 rs11591147 rs11206510 0.191 rs11591147 rs7523242 0.066 rs11591147 rs4927207 0.102 rs11591147 rs6662286 0.176 rs11591147 rs572512 0.008 rs11591147 rs1475701 0.028 rs11591147 rs7552841 0.004 rs1998013 rs11206510 0.191 rs1998013 rs7523242 0.066 rs1998013 rs4927207 0.102 rs1998013 rs6662286 0.176rs1998013 rs572512 0.008 rs1998013 rs1475701 0.028



Genetic variant 1 Genetic variant 2 Correlation rs1998013 rs7552841 0.004

rs11206510 rs7523242 0.186 rs11206510 rs4927207 0.028 rs11206510 rs6662286 0.017 rs11206510 rs572512 0.200 rs11206510 rs1475701 0.087 rs11206510 rs7552841 0.127 rs7523242 rs4927207 0.048 rs7523242 rs6662286 0.066 rs7523242 rs572512 0.409 rs7523242 rs1475701 0.057 rs7523242 rs7552841 0.068 rs4927207 rs6662286 0.201 rs4927207 rs572512 0.042 rs4927207 rs1475701 0.085 rs4927207 rs7552841 0.232 rs6662286 rs572512 0.163 rs6662286 rs1475701 0.059 rs6662286 rs7552841 0.025 rs572512 rs1475701 0.118rs572512 rs7552841 0.049 rs1475701 rs7552841 0.102



eTable 5. Burden of rare alleles in exome sequencing studies Burden of protein-truncating and missense variants predicted to be “probably deleterious” for protein function in 8,373 type 2 diabetes cases and 8,466 controls from exome sequencing studies.

Gene Class of genetic

variants

Carriers with type

2 diabetes

Non-carriers

with type 2

diabetes

Carriers among

controls

Non-carriers among

controls

Odds ratio of type 2 diabetes

for carriers (95% CI)

p-value

NPC1L1 Protein truncating 143 8230 129 8337

1.12 (0.88-1.43) 0.34

Probably deleterious missense

360 8013 294 8172 1.26

(1.07-1.47) 0.005

HMGCR Protein truncating 0 8373 0 8466 N/A N/A


3 8370 10 8456 0.31

(0.08-1.12) 0.07

PCSK9 Protein truncating 37 8336 33 8433

1.13 (0.71-1.82) 0.61


100 8273 85 8381 1.22

(0.91-1.64) 0.18

ABCG5 Protein truncating 5 8368 9 8457

0.59 (0.20-1.75) 0.34


54 8319 71 8395 0.77

(0.54-1.10) 0.15

ABCG8 Protein truncating 31 8342 35 8431

0.88 (0.55-1.44) 0.62


94 8279 112 8354 0.84

(0.64-1.11) 0.23

LDLR Protein truncating 2 8371 2 8464

1.02 (0.14-7.26) 0.98


53 8320 47 8419 1.15

(0.78-1.70) 0.49

N/A, not available (not calculated); CI, confidence interval.



eFigure 1. Meta-analysis results Meta-analysis of the association of LDL-cholesterol lowering polymorphisms with risk of type 2 diabetes in EPIC-InterAct,1 UK Biobank2 and DIAGRAM3. For rs12916 in HMGCR, results of an additional eleven studies reported by Swerdlow and colleagues13 were included. In EPIC-InterAct, genotyping was performed in two batches using the Illumina 660w quad and Illumina CoreExome genotyping arrays. Therefore, results of the main analysis are presented separately for individuals genotyped with the Illumina 660w quad array (InterAct-GWAS) and for individuals genotyped with the Illumina CoreExome array (InterAct-CoreExome). Squares indicate the odds ratios and error bars their 95% confidence interval. The size of the squares reflects the weight of the study in the inverse-variance weighted meta-analysis. OR indicates the odds ratio; CI, confidence interval.

1.8 1 1.25

Study OR (95% CI)

OR of type 2 diabetes per allele

NPC1L1 – rs2073547

InterAct-GWAS 1.11 (1.02, 1.21) 4,178

Cases Controls

4,254

InterAct-Core-Exome 1.05 (0.98, 1.13) 5,121 7,269

UK Biobank 1.03 (0.99, 1.08) 6,627 143,765

DIAGRAM 1.05 (1.02, 1.09) 34,840 114,981

Overall (I-squared = 0.0%, p = 0.48)

1.05 (1.03, 1.08) 50,775 270,269

1.8 1 1.25

NPC1L1 – rs217386

Study OR (95% CI) Cases Controls


InterAct-GWAS 1.03 (0.97, 1.10) 4,178 4,254


UK Biobank 1.01 (0.98, 1.05) 6,627 143,765

DIAGRAM 1.03 (1.01, 1.05) 34,840 114,981

Overall(I-squared = 0.0%, p = 0.68)

1.03 (1.01, 1.05) 50,775 270,269

1.25 .5 .8 1 1.25 2 4

HMGCR – rs12916



InterAct-GWAS 1.05 (0.98, 1.11) 4,178 4,254


UK Biobank 1.04 (1.00, 1.07) 6,627 143,765

DIAGRAM 1.01 (0.99, 1.04) 34,840 114,981


1.03 (1.01, 1.05) 55,271 320,946

WGHS 1.12 (1.04, 1.21) 1,444 21,268

ET2DS 1.25 (1.10, 1.42) 1,046 821

BWHHS 1.10 (0.95, 1.27) 438 2,839

WHII 1.01 (0.86, 1.19) 336 4,711

WHI 1.03 (0.87, 1.22) 282 5,427

JUPITER 0.83 (0.70, 0.98) 279 8,430

MESA 1.02 (0.83, 1.25) 220 2,078

NPHS-II 1.18 (0.96, 1.45) 217 2,449

CaPS 1.06 (0.81, 1.39) 118 1,288

CARDIA 1.15 (0.85, 1.56) 99 1,344

CFS 1.13 (0.48, 2.66) 17 22

1.8 1 1.25

HMGCR – rs5744707



1.06 (0.96, 1.18)InterAct-GWAS 4,178 4,254

0.94 (0.86, 1.03)InterAct-Core-Exome 5,121 7,269

0.98 (0.93, 1.04)UK Biobank 6,627 143,765

0.98 (0.95, 1.02)DIAGRAM 34,840 114,981

Overall (I-squared = 2.8%, p = 0.38)

0.98 (0.96, 1.01) 50,775 270,269

1.8 1 1.25

HMGCR – rs16872526



1.04 (0.93, 1.16)InterAct-GWAS 4,178 4,254


1.08 (1.02, 1.14)UK Biobank 6,627 143,765

0.99 (0.95, 1.03)DIAGRAM 34,840 114,981


1.02 (0.99, 1.05) 50,775 270,269

1.5 .8 1 1.25 2


PCSK9 – rs11591147


InterAct-GWAS 0.88 (0.67, 1.15) 4,178 4,254


UK Biobank 1.10 (0.97, 1.25) 6,627 143,765

DIAGRAM 1.09 (0.98, 1.22) 34,840 114,981


1.09 (1.01, 1.17) 50,775 270,269

1.8 1 1.25

ABCG5/G8 – rs4299376



0.96 (0.90, 1.03)InterAct-GWAS 4,178 4,254


1.03 (0.99, 1.07)UK Biobank 6,627 143,765

1.01 (0.98, 1.04)DIAGRAM 34,840 114,981


1.01 (0.99, 1.03) 50,775 270,269

1.8 1 1.25

LDLR – rs6511720



1.04 (0.94, 1.14)InterAct-GWAS 4,178 4,254


1.03 (0.98, 1.08)UK Biobank 6,627 143,765

1.02 (0.98, 1.06)DIAGRAM 34,840 114,981


1.03 (1.00, 1.06) 50,775 270,269



eFigure 2. Conditional analysis at the NPC1L1 locus

Association with LDL cholesterol at the NPC1L1 locus in the Global Lipids Genetics Consortium6 results before conditioning (left), after conditioning on the lead rs2073547 polymorphism (middle) and after conditioning on both the rs2073547 and rs217386 polymorphisms (right) in approximate conditional analyses using the GCTA software.14 After conditioning on two polymorphisms the signal was attenuated. Genomic coordinates are relative to Human Reference Genome Build 37.



eFigure 3. Conditional analysis at the PCSK9 locus

Panel A shows associations with LDL cholesterol at the PCSK9 locus before (left) and after (right) conditioning on rs11591147, rs1998013, rs11206510, rs7523242, rs4927207, rs6662286, rs572512, rs1475701, and rs7552841 in approximate conditional analyses using the GCTA software.14 Data are from the Global Lipids Genetics Consortium.6 After conditioning on the nine polymorphisms the signal was attenuated. Panel B shows associations with LDL cholesterol in a smaller sample with available individual level data. There was evidence of two distinct genome-wide significant signals (p<5 x 10-08) represented by rs11591147 and rs471705. The association signal in the region (left graph) was progressively attenuated after conditioning on rs11591147 (middle graph) and, then, after conditioning on both rs11591147 and rs471705 (right graph). Genomic coordinates are relative to Human Reference Genome Build 37.



eFigure 4. Associations of LDL-lowering alleles with continuous cardiometabolic traits Associations are in standardized units per LDL-cholesterol lowering allele. LDL cholesterol (N=173,021), HDL cholesterol (N=187,087), ln-transformed triglycerides (N=177,791) levels data are from the Global Lipids Genetics Consortium.6 Systolic (N=8,756) and diastolic (N=8,755) blood pressure data are from the EPIC-InterAct1 subcohort. Body mass index (N=333,495) and waist-to-hip ratio (N=224,047) data are from the GIANT Consortium9,10; fasting glucose (N=133,010), two hour glucose (N=42,854) and ln-transformed fasting insulin data (N=108,557) are from the MAGIC Consortium7,8. Abbreviations: LDL, low density lipoprotein cholesterol; HDL, high-density lipoprotein cholesterol; Ln-TG, triglycerides (natural logarithm transformed); SBP, systolic blood pressure; DBP, diastolic blood pressure; BMI, body mass index; WHR, waist-to-hip ratio; FPG, fasting plasma glucose; 2hr, 2 hour glucose; Ln-FI, fasting insulin (natural logarithm transformed); SD, standard deviation; CI, confidence interval. All genetic variants were strongly associated with LDL cholesterol levels. NPC1L1 polymorphisms were weakly associated with lower triglyceride levels, consistent with the effect of ezetimibe on triglyceride levels.15 HMGCR polymorphisms were associated with higher BMI levels, consistent with the effect of statins on body weight.13

NPC1L1 – rs2073547

HMGCR – rs16872526

NPC1L1 – rs217386

PCSK9 – rs11591147

HMGCR – rs12916

ABCG5/G8 – rs4299376

HMGCR – rs5744707

LDLR – rs6511720

2hr -0.01 (-0.03, 0.02)

Ln-TG -0.01 (-0.02, -0.00)

WHR -0.00 (-0.01, 0.01)

FPG 0.00 (-0.01, 0.01)

Ln-FI -0.01 (-0.02, 0.00)

DBP -0.02 (-0.05, 0.01)

SBP -0.02 (-0.04, 0.01)

HDL 0.00 (-0.01, 0.01)

BMI 0.00 (-0.00, 0.01)

LDL -0.04 (-0.04, -0.03)

0-.1 0 .1

FPG 0.00 (-0.01, 0.01)

Ln-FI 0.01 (0.00, 0.02)

DBP -0.00 (-0.03, 0.03)

WHR 0.01 (0.00, 0.01)

Ln-TG -0.00 (-0.01, 0.00)

LDL -0.07 (-0.08, -0.07)

HDL -0.00 (-0.01, 0.00)

2hr 0.01 (-0.00, 0.02)

SBP -0.02 (-0.04, 0.01)

BMI 0.02 (0.01, 0.02)

0-.1 0 .1

BMI -0.00 (-0.01, 0.01)

WHR -0.01 (-0.02, 0.01)

Ln-TG -0.01 (-0.02, 0.01)

DBP -0.03 (-0.08, 0.02)

LDL -0.05 (-0.07, -0.04)

SBP -0.03 (-0.07, 0.01)

FPG 0.01 (-0.01, 0.02)

Ln-FI -0.00 (-0.02, 0.02)

HDL 0.01 (-0.00, 0.02)

2hr 0.03 (-0.01, 0.06)

0-.1 0 .1

WHR 0.01 (-0.00, 0.02)

FPG 0.01 (-0.00, 0.02)

2hr -0.00 (-0.02, 0.02)

Ln-FI 0.01 (-0.00, 0.03)

LDL -0.22 (-0.23, -0.21)

Ln-TG -0.01 (-0.02, 0.00)

DBP 0.01 (-0.03, 0.06)

BMI 0.01 (-0.00, 0.02)

SBP 0.01 (-0.03, 0.06)

HDL 0.02 (0.01, 0.04)

0-.25 0 .25

Ln-FI -0.00 (-0.01, 0.01)

HDL 0.00 (-0.01, 0.01)

BMI 0.01 (0.00, 0.01)

WHR 0.00 (-0.00, 0.01)

SBP 0.01 (-0.02, 0.04)

LDL -0.08 (-0.09, -0.07)

FPG -0.00 (-0.01, 0.01)

DBP -0.01 (-0.04, 0.02)

Ln-TG -0.01 (-0.02, -0.00)

2hr -0.01 (-0.02, 0.01)

0-.1 0 .1

WHR 0.04 (0.01, 0.08)

FPG 0.04 (0.01, 0.07)

BMI 0.01 (-0.03, 0.04)

DBP 0.02 (-0.11, 0.14)

2hr 0.05 (-0.00, 0.10)

LDL -0.50 (-0.53, -0.46)

SBP 0.06 (-0.05, 0.17)

Ln-FI 0.03 (-0.02, 0.07)

HDL 0.04 (0.00, 0.07)

Ln-TG -0.01 (-0.04, 0.03)

0-.6 0 .6

LDL -0.04 (-0.05, -0.03)

BMI 0.02 (0.01, 0.03)

FPG 0.01 (-0.00, 0.03)

2hr 0.00 (-0.04, 0.04)

WHR 0.01 (-0.00, 0.02)

Ln-FI 0.01 (-0.01, 0.03)

Ln-TG 0.01 (-0.01, 0.02)

SBP 0.00 (-0.05, 0.05)

HDL -0.01 (-0.02, 0.00)

DBP 0.02 (-0.03, 0.08)

0-.1 0 .1

SBP 0.00 (-0.03, 0.04)

Ln-FI -0.00 (-0.02, 0.01)

2hr 0.01 (-0.02, 0.03)

FPG 0.01 (-0.00, 0.03)

WHR -0.00 (-0.01, 0.01)

BMI 0.00 (-0.01, 0.01)

DBP 0.00 (-0.04, 0.04)

Ln-TG -0.01 (-0.02, -0.01)

LDL -0.05 (-0.06, -0.04)

HDL 0.00 (-0.00, 0.01)

0-.1 0 .1

SD per allele SD per allele SD per allele SD per allele

SD per allele SD per allele SD per allele SD per allele

Phenotype Beta (95% CI) Phenotype Beta (95% CI) Phenotype Beta (95% CI) Phenotype Beta (95% CI)

Phenotype Beta (95% CI) Phenotype Beta (95% CI) Phenotype Beta (95% CI) Phenotype Beta (95% CI)



eFigure 5. Stratified associations of NPC1L1 variants Combined association of LDL-lowering alleles at NPC1L1 with risk of type 2 diabetes in strata of sex, age and body mass index. Data are from the EPIC-InterAct1 and UK Biobank2 studies. Analyses are scaled to represent the odds ratio of type 2 diabetes for a genetically predicted reduction in LDL cholesterol of 1 mmol/L. Squares indicate the odds ratio and the error bars its 95% confidence interval. The size of the squares indicates the weight of the subgroup analysis in the inverse-variance weighted meta-analysis. OR indicates the odds ratio; CI, confidence interval.

1.125 1 8 32

Body mass index

> 55 years

Age

Obese

Overweight

<= 55 years

All (I-squared = 48.4%, p = 0.144)

Men

All (I-squared = 0.0%, p = 0.939)

Lean

Sex

All (I-squared = 0.0%, p = 0.609)

Women

Analysis

1.17 (0.46, 2.97)

2.75 (1.07, 7.02)

2.89 (1.11, 7.53)

2.19 (1.19, 4.02)

2.17 (1.00, 4.71)

2.22 (1.25, 3.93)

2.37 (1.33, 4.22)

2.27 (0.97, 5.33)

2.11 (1.03, 4.35)

6.46 (1.40, 29.81)

OR (95% CI)

9,100

6,973

5,823

5,411

14,549

8,038

14,657

14,511

6,619

1,753

Cases

Odds ratio of type 2 diabetesper 1 mmol/L reduction

in LDL cholesterol

Controls

68,024

26,481

50,617

47,229

118,484

54,333

118,854

115,253

64,521

41,386



eFigure 6. Associations with continuous cardiometabolic traits Association of LDL lowering alleles with continuous anthropometric and glycemic traits. Associations are in standardised units per 1 mmol/L reduction in LDL cholesterol. Body mass index (N=333,495) and waist-to-hip ratio (N=224,047) data are from the GIANT Consortium9,10; fasting glucose (N=133,010), two hour glucose (N=42,854) and ln-transformed fasting insulin data (N=108,557) are from the MAGIC Consortium7,8. Abbreviations: BMI, body mass index; WHR, waist-to-hip ratio; FPG, fasting plasma glucose; 2hr, 2 hour glucose; Ln-FI, fasting insulin (natural logarithm transformed); CI, confidence interval.

BMI 0.08 (-0.04, 0.20)

0-.5 0 .5

0.23 (0.15, 0.30)

0-.5 0 .5

-0.02 (-0.09, 0.05)

0-.5 0 .5

0.09 (0.00, 0.17)

0-.5 0 .5

0.03 (-0.01, 0.08)

FPG 0.14 (-0.05, 0.33) 0.02 (-0.07, 0.11) 0.08 (0.02, 0.15) -0.02 (-0.11, 0.08) 0.04 (-0.01, 0.08)

Ln-FI -0.15 (-0.37, 0.07) 0.16 (0.05, 0.27) 0.05 (-0.03, 0.14) -0.04 (-0.16, 0.09) 0.05 (-0.01, 0.11)

2hr -0.05 (-0.51, 0.41) 0.14 (-0.03, 0.31) 0.10 (0.00, 0.20) -0.09 (-0.28, 0.11) -0.01 (-0.11, 0.09)

WHR -0.01 (-0.14, 0.12) 0.09 (0.01, 0.18) 0.08 (0.01, 0.15) 0.04 (-0.06, 0.14) 0.03 (-0.02, 0.08)

0-.5 0 .5

Standard deviationper 1 mmol/L genetically-predicted reduction

in LDL cholesterol

NPC1L1 HMGCR PCSK9 ABCG5/G8 LDLR

Beta (95% CI) Beta (95% CI) Beta (95% CI) Beta (95% CI) Beta (95% CI)

-1

Phenotype



eReferences 1. InterAct Consortium, Langenberg C, Sharp S, et al. Design and cohort description of the InterAct Project: an examination of the interaction of genetic

and lifestyle factors on the incidence of type 2 diabetes in the EPIC Study. Diabetologia. Sep 2011;54(9):2272-2282. 2. Collins R. What makes UK Biobank special? Lancet. Mar 31 2012;379(9822):1173-1174. 3. Morris AP, Voight BF, Teslovich TM, et al. Large-scale association analysis provides insights into the genetic architecture and pathophysiology of

type 2 diabetes. Nature genetics. Sep 2012;44(9):981-990. 4. Fuchsberger C, Flannick J, Teslovich TM, et al. The genetic architecture of type 2 diabetes. Nature. Jul 11 2016;536(7614):41-47. 5. Nikpay M, Goel A, Won HH, et al. A comprehensive 1,000 Genomes-based genome-wide association meta-analysis of coronary artery disease.

Nature genetics. Oct 2015;47(10):1121-1130. 6. Global Lipids Genetics Consortium, Willer CJ, Schmidt EM, et al. Discovery and refinement of loci associated with lipid levels. Nature genetics.

Nov 2013;45(11):1274-1283. 7. Scott RA, Lagou V, Welch RP, et al. Large-scale association analyses identify new loci influencing glycemic traits and provide insight into the

underlying biological pathways. Nature genetics. Sep 2012;44(9):991-1005. 8. Manning AK, Hivert MF, Scott RA, et al. A genome-wide approach accounting for body mass index identifies genetic variants influencing fasting

glycemic traits and insulin resistance. Nature genetics. Jun 2012;44(6):659-669. 9. Locke AE, Kahali B, Berndt SI, et al. Genetic studies of body mass index yield new insights for obesity biology. Nature. Feb 12 2015;518(7538):197-

206. 10. Shungin D, Winkler TW, Croteau-Chonka DC, et al. New genetic loci link adipose and insulin biology to body fat distribution. Nature. Feb 12

2015;518(7538):187-196. 11. Ference BA, Majeed F, Penumetcha R, Flack JM, Brook RD. Effect of naturally random allocation to lower low-density lipoprotein cholesterol on the

risk of coronary heart disease mediated by polymorphisms in NPC1L1, HMGCR, or both: a 2 x 2 factorial Mendelian randomization study. Journal of the American College of Cardiology. Apr 21 2015;65(15):1552-1561.

12. Johnson AD, Handsaker RE, Pulit SL, Nizzari MM, O'Donnell CJ, de Bakker PI. SNAP: a web-based tool for identification and annotation of proxy SNPs using HapMap. Bioinformatics. Dec 15 2008;24(24):2938-2939.

13. Swerdlow DI, Preiss D, Kuchenbaecker KB, et al. HMG-coenzyme A reductase inhibition, type 2 diabetes, and bodyweight: evidence from genetic analysis and randomised trials. Lancet. Jan 24 2015;385(9965):351-361.

14. Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: a tool for genome-wide complex trait analysis. American journal of human genetics. Jan 7 2011;88(1):76-82.

15. Cannon CP, Blazing MA, Giugliano RP, et al. Ezetimibe Added to Statin Therapy after Acute Coronary Syndromes. The New England journal of medicine. Jun 18 2015;372(25):2387-2397.


Documents

Supplementary Online Content - JAMA...© 2016 American Medical Association. All rights reserved. Supplementary Online Content Lotta LA, Sharp SJ, Burgess S, et al. Association between