Coffee consumption and the risk of rheumatoid arthritis and systemic lupus erythematosus: a Mendelian randomization study

Sang Cheol Bae, Young Ho Lee

Research output: Contribution to journalArticle

Abstract

We aimed to analyze the causal association between coffee consumption and the risk of rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE). We performed a two-sample Mendelian randomization (MR) analysis using the inverse-variance weighted (IVW), MR-Egger regression, and weighted median methods. We used publicly available summary statistics datasets of coffee consumption genome-wide association studies (GWASs) as an exposure variable and RA and SLE GWASs as outcomes. Four single-nucleotide polymorphisms (SNPs) from GWASs of coffee consumption were selected as instrumental variables (IVs) to improve inference: NCARD (rs16868941), POR (rs17685), CYP1A1 (rs2470893), and LAMB4 (rs382140). The IVW method showed a causal association between coffee consumption and RA (beta = 0.770, SE = 0.279, p = 0.006). MR-Egger regression revealed that directional pleiotropy was unlikely to be biasing the result (intercept = − 0.145, p = 0.451). While the MR-Egger analysis showed no causal association between coffee consumption and RA (beta = 2.744, SE = 1.712, p = 0.355), the weighted median approach demonstrated a causal association between coffee consumption and RA (beta = 0.751, SE = 0.348, p = 0.031). However, the associations based on the weighted median analyses after the Bonferroni correction were not significant (adjusted p values = 0.091). The IVW, MR-Egger analysis, and weighted median methods showed no causal association between coffee consumption and SLE risk (beta = 0.594, SE = 0.437, p = 0.209; beta = 3.100, SE = 3.632, p = 0.550; beta = 0.733, SE = 0.567, p = 0.196). MR analysis results do not support causal associations between coffee consumption and the development of RA and SLE.

Original languageEnglish
Pages (from-to)2875-2879
Number of pages5
JournalClinical Rheumatology
Volume37
Issue number10
DOIs
Publication statusPublished - 2018 Oct 1

Fingerprint

Coffee
Random Allocation
Systemic Lupus Erythematosus
Mendelian Randomization Analysis
Rheumatoid Arthritis
Genome-Wide Association Study
Cytochrome P-450 CYP1A1
Single Nucleotide Polymorphism

Keywords

  • Coffee
  • Mendelian randomization
  • RA
  • SLE

ASJC Scopus subject areas

  • Rheumatology

Cite this

Coffee consumption and the risk of rheumatoid arthritis and systemic lupus erythematosus : a Mendelian randomization study. / Bae, Sang Cheol; Lee, Young Ho.

In: Clinical Rheumatology, Vol. 37, No. 10, 01.10.2018, p. 2875-2879.

Research output: Contribution to journalArticle

@article{9378b116ab0d4636bc4b96a1c2acc8b7,
title = "Coffee consumption and the risk of rheumatoid arthritis and systemic lupus erythematosus: a Mendelian randomization study",
abstract = "We aimed to analyze the causal association between coffee consumption and the risk of rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE). We performed a two-sample Mendelian randomization (MR) analysis using the inverse-variance weighted (IVW), MR-Egger regression, and weighted median methods. We used publicly available summary statistics datasets of coffee consumption genome-wide association studies (GWASs) as an exposure variable and RA and SLE GWASs as outcomes. Four single-nucleotide polymorphisms (SNPs) from GWASs of coffee consumption were selected as instrumental variables (IVs) to improve inference: NCARD (rs16868941), POR (rs17685), CYP1A1 (rs2470893), and LAMB4 (rs382140). The IVW method showed a causal association between coffee consumption and RA (beta = 0.770, SE = 0.279, p = 0.006). MR-Egger regression revealed that directional pleiotropy was unlikely to be biasing the result (intercept = − 0.145, p = 0.451). While the MR-Egger analysis showed no causal association between coffee consumption and RA (beta = 2.744, SE = 1.712, p = 0.355), the weighted median approach demonstrated a causal association between coffee consumption and RA (beta = 0.751, SE = 0.348, p = 0.031). However, the associations based on the weighted median analyses after the Bonferroni correction were not significant (adjusted p values = 0.091). The IVW, MR-Egger analysis, and weighted median methods showed no causal association between coffee consumption and SLE risk (beta = 0.594, SE = 0.437, p = 0.209; beta = 3.100, SE = 3.632, p = 0.550; beta = 0.733, SE = 0.567, p = 0.196). MR analysis results do not support causal associations between coffee consumption and the development of RA and SLE.",
keywords = "Coffee, Mendelian randomization, RA, SLE",
author = "Bae, {Sang Cheol} and Lee, {Young Ho}",
year = "2018",
month = "10",
day = "1",
doi = "10.1007/s10067-018-4278-9",
language = "English",
volume = "37",
pages = "2875--2879",
journal = "Clinical Rheumatology",
issn = "0770-3198",
publisher = "Springer London",
number = "10",

}

TY - JOUR

T1 - Coffee consumption and the risk of rheumatoid arthritis and systemic lupus erythematosus

T2 - a Mendelian randomization study

AU - Bae, Sang Cheol

AU - Lee, Young Ho

PY - 2018/10/1

Y1 - 2018/10/1

N2 - We aimed to analyze the causal association between coffee consumption and the risk of rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE). We performed a two-sample Mendelian randomization (MR) analysis using the inverse-variance weighted (IVW), MR-Egger regression, and weighted median methods. We used publicly available summary statistics datasets of coffee consumption genome-wide association studies (GWASs) as an exposure variable and RA and SLE GWASs as outcomes. Four single-nucleotide polymorphisms (SNPs) from GWASs of coffee consumption were selected as instrumental variables (IVs) to improve inference: NCARD (rs16868941), POR (rs17685), CYP1A1 (rs2470893), and LAMB4 (rs382140). The IVW method showed a causal association between coffee consumption and RA (beta = 0.770, SE = 0.279, p = 0.006). MR-Egger regression revealed that directional pleiotropy was unlikely to be biasing the result (intercept = − 0.145, p = 0.451). While the MR-Egger analysis showed no causal association between coffee consumption and RA (beta = 2.744, SE = 1.712, p = 0.355), the weighted median approach demonstrated a causal association between coffee consumption and RA (beta = 0.751, SE = 0.348, p = 0.031). However, the associations based on the weighted median analyses after the Bonferroni correction were not significant (adjusted p values = 0.091). The IVW, MR-Egger analysis, and weighted median methods showed no causal association between coffee consumption and SLE risk (beta = 0.594, SE = 0.437, p = 0.209; beta = 3.100, SE = 3.632, p = 0.550; beta = 0.733, SE = 0.567, p = 0.196). MR analysis results do not support causal associations between coffee consumption and the development of RA and SLE.

AB - We aimed to analyze the causal association between coffee consumption and the risk of rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE). We performed a two-sample Mendelian randomization (MR) analysis using the inverse-variance weighted (IVW), MR-Egger regression, and weighted median methods. We used publicly available summary statistics datasets of coffee consumption genome-wide association studies (GWASs) as an exposure variable and RA and SLE GWASs as outcomes. Four single-nucleotide polymorphisms (SNPs) from GWASs of coffee consumption were selected as instrumental variables (IVs) to improve inference: NCARD (rs16868941), POR (rs17685), CYP1A1 (rs2470893), and LAMB4 (rs382140). The IVW method showed a causal association between coffee consumption and RA (beta = 0.770, SE = 0.279, p = 0.006). MR-Egger regression revealed that directional pleiotropy was unlikely to be biasing the result (intercept = − 0.145, p = 0.451). While the MR-Egger analysis showed no causal association between coffee consumption and RA (beta = 2.744, SE = 1.712, p = 0.355), the weighted median approach demonstrated a causal association between coffee consumption and RA (beta = 0.751, SE = 0.348, p = 0.031). However, the associations based on the weighted median analyses after the Bonferroni correction were not significant (adjusted p values = 0.091). The IVW, MR-Egger analysis, and weighted median methods showed no causal association between coffee consumption and SLE risk (beta = 0.594, SE = 0.437, p = 0.209; beta = 3.100, SE = 3.632, p = 0.550; beta = 0.733, SE = 0.567, p = 0.196). MR analysis results do not support causal associations between coffee consumption and the development of RA and SLE.

KW - Coffee

KW - Mendelian randomization

KW - RA

KW - SLE

UR - http://www.scopus.com/inward/record.url?scp=85053288759&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85053288759&partnerID=8YFLogxK

U2 - 10.1007/s10067-018-4278-9

DO - 10.1007/s10067-018-4278-9

M3 - Article

C2 - 30167974

AN - SCOPUS:85053271673

VL - 37

SP - 2875

EP - 2879

JO - Clinical Rheumatology

JF - Clinical Rheumatology

SN - 0770-3198

IS - 10

ER -