Figure 1.

Directed acyclic graph illustrating the plausible (a) or known (b) relationships among key variables [solid arrows; e.g. Mendelian randomization (MR) analyses show a causal effect of higher urate resulting in higher systolic blood pressure (SBP) and higher SBP resulting in lower birthweight] where offspring birthweight is the outcome and maternal urate is the exposure of interest. (a) The putative relationships of the main model we aimed to investigate, estimating the total causal effect of maternal urate on offspring birthweight where the asterisk denotes the causal association being tested. Multivariable observational analyses suggest an inverse effect of higher maternal circulating urate lowering infant birthweight. We used MR with maternal genetic instrumental variables (G_M) to explore whether there was an inverse total causal effect of urate on birthweight and, if so, whether urate might be a confounder of the SBP–birthweight inverse effect. A rectangular box denotes the variable and its effects have been adjusted for. As offspring genome (G_Off) has effects on own birthweight, maternal effects were estimated as independent of the offspring genome to capture the maternal genome’s effect on offspring birthweight alone. (b) The known relationships between urate and SBP, and SBP and offspring birthweight. SBP, systolic blood pressure; G_M, maternal genotype; G_Off, offspring’s genotype