Abstract and Introduction
Abstract
Background & Aims: The causal association of lower birthweight with non-alcoholic fatty liver disease (NAFLD) and the mediating pathways remain unclear. We aimed to investigate the causal, independent association of lower birthweight with NAFLD and identify potential metabolic mediators and their mediation effects in this association.
Methods: We performed two-step, two-sample Mendelian randomization (MR) using genome-wide association study (GWAS) summary statistics for birthweight from the Early Growth Genetics Consortium of 298 142 Europeans, NAFLD from a GWAS meta-analysis of 8434 NAFLD cases and 770 180 controls of Europeans, and 25 candidate mediators from corresponding reliable GWASs.
Results: Genetically determined each 1-SD lower birthweight was associated with a 45% (95% CI: 1.25–1.69) increased risk of NAFLD, and this causal association persisted after adjusting for childhood obesity or adult adiposity traits in multivariable MR. Two-step MR identified 6 of 25 candidate mediators partially mediate the effect of lower birthweight on NAFLD, including fasting insulin (proportion mediated: 22.05%), leucine (17.29%), isoleucine (13.55%), valine (11.37%), alanine (10.01%) and monounsaturated fatty acids (MUFA; 7.23%). Bidirectional MR suggested a unidirectional effect of insulin resistance on isoleucine, leucine and valine and a unidirectional effect of alanine on insulin resistance.
Conclusions: This MR study elucidated the causal impact of lower birthweight on subsequent risk of NAFLD, independently of later-life adiposity and identified mediators including insulin resistance, branched-chain amino acids, alanine and MUFA in this association pathway. Our findings shed light on the pathogenesis of NAFLD and imply additional targets for prevention and intervention of NAFLD attributed to low birthweight.
Introduction
Non-alcoholic fatty liver disease (NAFLD) is the principal aetiology contributing to advanced liver diseases such as cirrhosis and hepatocellular carcinoma and has become a major public health threat affecting more than 25% of the global population.[1,2] Because obesity and type 2 diabetes are the most important risk factors for NAFLD, the burden of NAFLD is expected to increase at an alarming rate as the epidemics of obesity and diabetes continue to grow.[3] According to the Developmental Origins of Health and Disease hypothesis, early-life environmental exposures (e.g., intrauterine growth restriction) during the prenatal development period can permanently change the body's structure, physiology and metabolism, influencing the susceptibility to cardiometabolic disease, primarily type 2 diabetes, in later life.[4,5] Intrauterine growth restriction affects the pancreas, adipose tissue and liver, which are postulated to programme for hepatic insulin resistance and subsequent NAFLD,[5,6] therefore, the developmental origin is legitimate for NAFLD.
Thus far, evidence for the impact of intrauterine growth restriction on NAFLD in later life is limited.[6] One recent population-based observational study associated low birthweight, a widely used indicator of retarded fetal growth and intrauterine malnutrition, with an increased risk of NAFLD in adulthood, but this association was largely diminished after adjusting for metabolic impairment, implying that low birthweight may not increase the risk of NAFLD per se, but rather through metabolic disorders.[6] However, because conventional observational studies are vulnerable to residual confounding, it remains unclear whether lower birthweight has a causal effect on the development of NAFLD. Evidence in humans and non-human primates suggests that intrauterine growth restriction may lead to persistent changes in insulin sensitivity and hepatic metabolism of glucose, fatty acids and amino acids, producing metabolic clues that prime the liver for NAFLD and increase susceptibility to NASH.[5] In addition, there are impairments in amino acid metabolism associated with insulin resistance, as well as alterations in fatty acids and intact lipids, which are partially because of altered de novo liver lipogenesis, lipolysis rate and very low-density lipoprotein metabolism in NAFLD.[7–9] Elucidating whether these metabolic changes mediate the relationship between lower birthweight and NAFLD will improve the understanding of the pathophysiology of NAFLD and provide additional opportunities for prevention and intervention to curb the NAFLD epidemic. To date, little attention has been paid to the mediating pathway between lower birthweight and NAFLD.
To fill the knowledge gap, we performed Mendelian randomization (MR) analyses to investigate the causal, independent association between lower birthweight and NAFLD risk, with particular interests in identifying potential mediators and quantifying their mediation effects in this association pathway. MR approaches using genetic variants as instrumental variables (IVs) to infer causality between related traits. Since genetic variation is randomly assigned at conception, MR analyses are analogous to randomized clinical trials and are less susceptible to confounding and reverse causality.[10] For mediation analysis, two-step MR is both sensitive to the causal effects of the mediator and corrects for its measurement error.[11]
Liver International. 2023;43(4):829-839. © 2023 Blackwell Publishing
