Mendelian randomization analysis rules out disylipidaemia as colorectal cancer cause

Gemma Ibáñez‐Sanz ,Nieves Ascunce ,Javier Llorca ,Marina Pollán ,Camilo Palazuelos ,Gemma Castaño‐Vinyals ,Adonina Tardón ,José Juan Jiménez Moleón ,Pablo Navarro ,Manolis Kogevinas ,Valle Olmos Juste ,Beatriz Pérez‐Gómez ,Guillermo Fernández‐Tardón ,Anna Díez-Villanueva ,Mariona Bustamante ,Tania Fernández‐Villa ,Inmaculada Salcedo‐Bellido ,Marina Riera-Ponsati ,Dolores Salas ,Marta Crous‐Bou ,Luis Ortega-Valín ,Núria Aragonés ,Pilar Amiano ,Eva Ardanáz ,Rocio Alvarez ,Jone M Altzibar ,Vı́ctor Moreno

doi:10.1038/s41598-019-49880-w

Abstract

Dyslipidemia and statin use have been associated with colorectal cancer (CRC), but prospective studies have shown mixed results. We aimed to determine whether dyslipidemia is causally linked to CRC risk using a Mendelian randomization approach and to explore the association of statins with CRC. A case-control study was performed including 1336 CRC cases and 2744 controls (MCC-Spain). Subjects were administered an epidemiological questionnaire and were genotyped with an array which included polymorphisms associated with blood lipids levels, selected to avoid pleiotropy. Four genetic lipid scores specific for triglycerides (TG), high density lipoprotein cholesterol (HDL), low density lipoprotein cholesterol (LDL), or total cholesterol (TC) were created as the count of risk alleles. The genetic lipid scores were not associated with CRC. The ORs per 10 risk alleles, were for TG 0.91 (95%CI: 0.72–1.16, p = 0.44), for HDL 1.14 (95%CI: 0.95–1.37, p = 0.16), for LDL 0.97 (95%CI: 0.81–1.16, p = 0.73), and for TC 0.98 (95%CI: 0.84–1.17, p = 0.88). The LDL and TC genetic risk scores were associated with statin use, but not the HDL or TG. Statin use, overall, was a non-significant protective factor for CRC (OR 0.84; 95%CI: 0.70–1.01, p = 0.060), but lipophilic statins were associated with a CRC risk reduction (OR 0.78; 95%CI 0.66–0.96, p = 0.018). Using the Mendelian randomization approach, our study does not support the hypothesis that lipid levels are associated with the risk of CRC. This study does not rule out, however, a possible protective effect of statins in CRC by a mechanism unrelated to lipid levels.

Highlights

Mendelian randomization is a technique used to determine the causal impact of a risk factor on an outcome from observational data using genetic variants
In our study, using the Mendelian randomization approach, none of the lipid genetic scores for dyslipidemia analyzed was associated with Colorectal cancer (CRC) risk
This indicates that lifetime dyslipidemia most probably is unrelated to the development of colorectal neoplasms, and that the associations previously found in observational studies between dyslipidemia and CRC could be result of uncontrolled confounding factors or reverse causation

Summary

Introduction

Findings from prospective studies that have examined the association between serum dyslipidemia (TG, HDL, low density lipoprotein cholesterol (LDL) or total cholesterol (TC)) and colorectal neoplasia have been inconsistent[6,8,9,10,11] It is unknown whether lipids and lipoproteins cause cancer or are intermediate or correlated factors within carcinogenic pathways. A previous Mendelian randomization analysis assessing the causality of dyslipidemia and CRC has been published before[11] It reported an association between a genetic score for TC and the risk of CRC (OR per unit SD increase = 1.46; 95% CI: 1.20–1.79) but no significant association was found for LDL, HDL or TG. Epidemiological studies on dyslipidemia and CRC risk could be confounded by 3-Hydroxy3-methylglutaryl-coenzyme A reductase inhibitors (statins) use, which might have a protective effect on CRC. We wanted to explore the effect of statins use on CRC

Objectives

Methods

Results

Conclusion