Research

Mixed

Mendelian randomization using semiparametric methods (fractional polynomial or piecewise linear) reveals that the causal relationship between BMI and systolic blood pressure is nonlinear, showing a steep increase up to a BMI of ~32 kg/m2 and a flatter or slightly negative slope thereafter, while the relationship with diastolic blood pressure increases almost linearly up to ~40 kg/m2.

For individuals with a BMI between 25 and 32, blood pressure tends to rise sharply with each unit of BMI. However, once BMI exceeds 32, the rate of blood pressure increase may slow down or even slightly decrease, suggesting a complex biological interaction rather than a simple linear progression. For Diastolic Blood Pressure, the increase remains relatively linear up to a BMI of 40. This implies that weight management strategies might have different impacts on blood pressure depending on the starting BMI and the type of blood pressure being monitored.

GoodQualifiesHIGH confidence
The estimate of the exposure-outcome relationship from the piecewise linear method visually suggested a threshold–type relationship, with a steep slope up to a BMI value of about 32 kg/m2, and a slightly negative slope from 32 kg/m2 onwards... The relationship between BMI and DBP was also nonlinear... the exposure-outcome curve increasing almost linearly up to a BMI of around 40 kg/m2.
James R Staley et al. · Genetic Epidemiology · 2017

Why this rating

Large sample size (UK Biobank, N=133,687), use of robust instrumental variable methods, and consistent findings across multiple statistical tests, though observational in nature despite MR design.

Source

Semiparametric methods for estimation of a nonlinear exposure‐outcome relationship using instrumental variables with application to Mendelian randomization

James R Staley et al. · Genetic Epidemiology · 2017

DOI 10.1002/gepi.22041

mechanism_only · n=502682Cited 312×
Read the paper
DOI resolved against Crossref · corpus check 2026-06-10

This is one finding among thousands. Every one is graded and traced to its source, so you can see what the evidence actually supports. Browse the research →