BMI used to look like it protected people who were overweight—but when you look at belly fat, that protection disappears, and higher BMI just keeps getting riskier.
Scientific Claim
The relationship between BMI and all-cause mortality shifts from J-shaped to linearly negative after adjusting for waist circumference, indicating that BMI’s apparent protective effect at higher weights is largely due to unmeasured central adiposity.
Original Statement
“the RCS analyses for BMI changed from a J-shaped (P < 0.05 for non-linearity test) to a negative association (P < 0.01)”
Evidence Quality Assessment
Claim Status
appropriately stated
Study Design Support
Design supports claim
Appropriate Language Strength
association
Can only show association/correlation
Assessment Explanation
The study uses RCS to model non-linearity and reports statistical significance of the shift after adjustment; language appropriately reflects observational association.
Gold Standard Evidence Needed
According to GRADE and EBM methodology, here is what ideal scientific evidence would look like to definitively prove or disprove this specific claim, ordered from strongest to weakest evidence.
Systematic Review & Meta-AnalysisLevel 1aIn EvidenceWhether adjusting for central adiposity consistently eliminates the J-shaped BMI-mortality curve across global studies.
Whether adjusting for central adiposity consistently eliminates the J-shaped BMI-mortality curve across global studies.
What This Would Prove
Whether adjusting for central adiposity consistently eliminates the J-shaped BMI-mortality curve across global studies.
Ideal Study Design
A meta-analysis of 30+ prospective cohorts using both BMI and WC, modeling non-linear associations with RCS, and comparing BMI-mortality curves before and after WC adjustment.
Limitation: Cannot determine if the shift is causal or due to residual confounding.
Prospective Cohort StudyLevel 2bWhether the J-shaped curve is driven by undiagnosed illness in normal-weight individuals or by central fat in overweight individuals.
Whether the J-shaped curve is driven by undiagnosed illness in normal-weight individuals or by central fat in overweight individuals.
What This Would Prove
Whether the J-shaped curve is driven by undiagnosed illness in normal-weight individuals or by central fat in overweight individuals.
Ideal Study Design
A cohort of 8,000 adults with annual BMI, WC, and biomarkers of subclinical disease (CRP, HbA1c, albumin), followed for 20 years, testing mediation of BMI-mortality by illness and WC.
Limitation: Cannot prove causality or intervention effect.
Randomized Controlled TrialLevel 1bWhether reducing central fat in overweight individuals eliminates the mortality risk associated with higher BMI.
Whether reducing central fat in overweight individuals eliminates the mortality risk associated with higher BMI.
What This Would Prove
Whether reducing central fat in overweight individuals eliminates the mortality risk associated with higher BMI.
Ideal Study Design
A 10-year RCT of 2,000 adults with BMI 25–35 and high WC, randomized to intensive weight loss targeting central fat vs. control, with mortality as primary endpoint.
Limitation: Ethical and logistical barriers to long-term RCTs on mortality.
Evidence from Studies
Supporting (1)
When scientists looked at body weight (BMI) alone, heavier people seemed to live longer — but when they also measured belly fat, that ‘protective’ effect vanished. This means it wasn’t the extra weight that helped — it was the lack of belly fat.