Distinct pathway-based effects of blood pressure and body mass index on cardiovascular traits: comparison of novel Mendelian randomization approaches

blood pressure
body mass index
genetic heterogeneity
mendelian randomization
pathway analysis
tissue specificity
  • Objective: This study introduced a novel pathway-partitioned Mendelian randomization (MR) approach (based on proximity to Mendelian disease genes) and compared its findings to the existing tissue-partitioned MR approach for Blood Pressure (BP) and Body Mass Index (BMI) effects on cardiovascular traits.
  • BP Findings: The pathway-partitioned MR showed that renal-system related IVs were significantly associated with Coronary Artery Disease (CAD), while vascular-related IVs were significantly associated with Stroke, demonstrating distinct causal pathways.
  • BMI Findings: Both the pathway-partitioned (Mental Health vs. Metabolic) and tissue-partitioned (Brain vs. Adipose) methods for BMI consistently found that genetic variants related to central/appetite regulation had a stronger causal effect on the risk of Type 2 Diabetes Mellitus (T2DM).
Published

23 January 2026

PubMed: 40375348 DOI: 10.1186/s13073-025-01472-2 Overview generated by: Gemini 2.5 Flash, 28/11/2025

Key Findings: Comparing MR Partitioning Approaches

This study introduces a novel pathway-partitioned Mendelian randomization (MR) approach and compares its performance and findings against an existing tissue-partitioned MR approach to dissect the heterogeneous genetic effects of Blood Pressure (BP) and Body Mass Index (BMI) on downstream cardiovascular traits.

  • Pathway-Partitioned MR is Effective: The novel pathway-partitioned approach successfully isolated sets of genetic instruments based on their proximity to genes implicated in Mendelian disorders of the renal system or vasculature (for BP), or mental health or metabolic disorders (for BMI).
  • Distinct Causal Pathways for BP:
    • Renal Pathway: Genetic instruments related to the renal system were significantly and strongly associated with a causal risk of Coronary Artery Disease (CAD).
    • Vascular Pathway: Genetic instruments related to the vasculature were significantly and strongly associated with a causal risk of Stroke.
  • Confirmation of BMI Pathways: The pathway-partitioned MR for BMI (Mental Health vs. Metabolic disorders) largely confirmed the findings of the previously developed tissue-partitioned approach (Brain vs. Adipose tissue). Both methodologies consistently showed that genetic variants related to central/appetite regulation (Mental Health/Brain-mediated) had a stronger causal effect on Type 2 Diabetes Mellitus (T2DM) risk compared to the metabolic/peripheral pathways.
  • Complementary Methods: The study concludes that both the pathway-partitioned and tissue-partitioned MR methods are valid, complementary tools for disentangling the genetic heterogeneity of complex traits, offering more granular insights into disease mechanisms than conventional MR which uses a single, aggregated instrument set.

Study Design and Methods

Study Design

The study employed a two-sample Mendelian randomization (MR) design, introducing a new method for partitioning genetic instrumental variables (IVs) and comparing it against a previous method (tissue-partitioned MR). The analysis was conducted using both individual-level (UK Biobank) and summary-level MR methodologies.

Instrument Partitioning Approaches

  1. Pathway-Partitioned MR (Novel Method):
    • Mechanism: SNPs were partitioned based on their proximity to genes associated with specific Mendelian diseases known to affect the trait.
    • BP Partitioning: IVs were grouped into Renal System (e.g., genes causing monogenic hypertension/kidney disease) and Vasculature (e.g., genes causing monogenic vascular disorders).
    • BMI Partitioning: IVs were grouped into Mental Health (genes implicated in psychiatric/neurodevelopmental disorders, often affecting appetite) and Metabolic disorders (genes linked to monogenic obesity/metabolic conditions).
  2. Tissue-Partitioned MR (Existing Method):
    • Mechanism: SNPs were partitioned based on genetic colocalization and causal association with gene expression in contrasting tissues.
    • BMI Partitioning: IVs were grouped into Brain-mediated and Adipose-mediated sets, mirroring the previous study by the same authors.

Data Sources and Outcomes

  • Exposure GWAS: Summary statistics for BP and BMI were sourced from large, publicly available consortia (e.g., UK Biobank, GIANT).
  • Outcome GWAS: Cardiometabolic outcomes included CAD, Stroke (all subtypes), and T2DM.

Statistical Analysis

  • MR Methods: The Inverse-Variance Weighted (IVW) method was the primary technique for estimating causal effects within each partitioned gene set. This was supplemented by sensitivity analyses (e.g., MR-Egger, Weighted Median).
  • Comparison: The causal effect estimates (e.g., Odds Ratios, ORs) derived from the different pathway/tissue partitions were compared using formal statistical tests to determine if the effect estimates were significantly different from each other.

Conclusions and Recommendations

This research validates the use of partitioning complex trait instruments into biologically meaningful subgroups to gain clearer etiological insights.

  • Enhanced Mechanistic Detail: By distinguishing between genetic variants acting on the renal system versus the vasculature for BP, the study provides targets for both CAD and Stroke prevention that were not possible with an aggregated BP instrument.
  • Therapeutic Prioritization: The consistent finding across both partitioning methods for BMI—that appetite-regulating pathways have a disproportionate effect on T2DM risk—strongly supports therapeutic targeting of central regulation for T2DM prevention in at-risk individuals.
  • Future Application: The authors propose that the pathway-partitioned approach, being data-driven and leveraging established Mendelian disease classifications, offers a valuable, mechanistic complement to tissue-based MR for dissecting the genetic architecture of other complex diseases.