Fetal genome predicted birth weight and polycystic ovary syndrome in later life: a Mendelian randomization study

Abstract

Associations between lower birth weight and higher polycystic ovary syndrome (PCOS) risk have been reported in previous observational studies, however, the causal relationship is still unknown. Based on decomposed fetal and maternal genetic effects on birth weight (n = 406,063), we conducted a two-sample Mendelian randomization (MR) analysis to assess potential causal relationships between fetal genome predicted birth weight and PCOS risk using a large-scale genome-wide association study (GWAS) including 4,138 PCOS cases and 20,129 controls. To further eliminate the maternally transmitted or non-transmitted effects on fetal growth, we performed a secondary MR analysis by utilizing genetic instruments after excluding maternally transmitted or non-transmitted variants, which were identified in another birth weight GWAS (n = 63,365 parent-offspring trios from Icelandic birth register). Linkage disequilibrium score regression (LDSR) analysis was conducted to estimate the genetic correlation. We found little evidence to support a causal effect of fetal genome determined birth weight on the risk of developing PCOS (primary MR analysis, OR: 0.86, 95% CI: 0.52 to 1.43; secondary MR analysis, OR: 0.86, 95% CI: 0.54 to 1.39). In addition, a marginally significant genetic correlation (r_g = -0.14, se = 0.07) between birth weight and PCOS was revealed via LDSR analysis. Our findings indicated that observed associations between birth weight and future PCOS risk are more likely to be attributable to genetic pleiotropy driven by the fetal genome rather than a causal mechanism.

Keywords: Mendelian randomization; birth weight; fetal genome; genetic pleiotropy; polycystic ovary syndrome.

Figure 1

Study design of MR analyses. (A) rs560887 and rs10872678 were identified as maternally transmitted and non-transmitted alleles respectively in the birth weight GWAS by Juliusdottir et al. (30). (B) SNPs were genome-wide significantly associated with potential confounders of PCOS, including BMI, type 2 diabetes, waist/hip circumference, waist-to-hip ratio, metabolic syndrome, glucose metabolism, and lipid metabolism. BMI, body mass index; BW, birth weight; EstBB, Estonian Biobank; GWAS, genome-wide association study; PCOS, polycystic ovary syndrome; SNP, single nucleotide polymorphism; WHR, waist-to-hip ratio.

Figure 2

Causal effects of fetal genome determined birth weight on future PCOS risk estimated in the primary MR analysis. Squares represent ORs of PCOS per SD increase in birth weight. Error bars represent 95% confidence intervals. A. 10 SNPs that were genome-wide significantly associated with potential confounders of PCOS, including BMI, type 2 diabetes, waist/hip circumference, waist-to-hip ratio, metabolic syndrome, glucose metabolism, and lipid metabolism, were excluded from the MR analysis. BMI, body mass index; CI, confidence interval; IVs, instrumental variables; IVW, inverse variance weighted; MR, Mendelian randomization; MR-PRESSO, Mendelian Randomization Pleiotropy RESidual Sum and Outlier; OR, odds ratio; P, p-value; PCOS, polycystic ovary syndrome; SD, standard deviation; SNP, single nucleotide polymorphism.

Figure 3

Causal effects of fetal genome determined birth weight on future PCOS risk estimated in the secondary MR analysis. Squares represent ORs of PCOS per SD increase in birth weight. Error bars represent 95% confidence intervals. (A) rs560887 and rs10872678were identified as maternally transmitted and non-transmitted alleles respectively in the birth weight GWAS by Juliusdottir et al. (30). (B) 9 SNPs that were genome-wide significantly associated with potential confounders of PCOS, including BMI, type 2 diabetes, waist/hip circumference, waist-to-hip ratio, metabolic syndrome, glucose metabolism, and lipid metabolism, were excluded from the MR analysis. BMI, body mass index; CI, confidence interval; IVs, instrumental variables; IVW, inverse variance weighted; MR, Mendelian randomization; MR-PRESSO, Mendelian Randomization Pleiotropy RESidual Sum and Outlier; OR, odds ratio; P, p-value; PCOS, polycystic ovary syndrome; SD, standard deviation; SNP, single nucleotide polymorphism.