Body composition conversations tend to fixate on the wrong number. Total body fat percentage is the metric most people anchor on, partly because it's the one that appears on smart scales and DEXA reports in big print. It correlates loosely with health outcomes. It does not predict them well.
The metric that does predict metabolic outcomes — insulin resistance, cardiovascular risk, hepatic steatosis, all-cause mortality — is visceral adipose tissue, or VAT. Two people at the same total body fat percentage can have a threefold difference in visceral fat. The one with more VAT has dramatically worse metabolic markers, even if they look indistinguishable on a tape measure.
Subcutaneous vs visceral — why the location matters
Adipose tissue is not a single thing. Subcutaneous fat — the layer under the skin — is metabolically relatively quiet. It stores energy, releases it slowly, and produces hormones (adiponectin, leptin) that mostly support metabolic health. Excess subcutaneous fat is cosmetically noticeable but, in moderate quantities, metabolically benign.
Visceral fat is different. It surrounds the organs of the abdominal cavity — liver, pancreas, intestines — and it is hormonally active in ways that drive disease. VAT secretes pro-inflammatory cytokines (TNF-α, IL-6) into the portal circulation, which goes directly to the liver. It releases free fatty acids continuously, which the liver converts into triglycerides and exports as VLDL particles. It interferes with insulin signalling at multiple tissue sites.
The same kilogram of fat, depending on where it's deposited, produces very different metabolic consequences. A person with 25kg of subcutaneous fat and 1kg of visceral fat is metabolically healthier than a person with 12kg of subcutaneous fat and 4kg of visceral fat — despite the second person being lighter and visually leaner.
What the research actually shows
The 2019 INTERHEART study, looking at heart attack risk across 52 countries, found that waist-to-hip ratio — a rough proxy for visceral fat distribution — was a stronger predictor of myocardial infarction than BMI in every population studied. The relative risk associated with the highest waist-to-hip ratio quintile was 2.5 times that of the lowest, after adjusting for BMI, age, sex, and smoking.
A 2021 meta-analysis in the Journal of the American College of Cardiology followed over 50,000 individuals with DEXA-measured visceral fat for an average of 10 years. Each one-standard-deviation increase in VAT was associated with a 35% increased risk of cardiovascular events, independent of BMI and total fat mass. Subcutaneous fat showed no significant relationship after adjustment.
The practical implication: Standing on the scale every morning, watching BMI, even tracking total body fat percentage — these are not bad metrics, but they are not the metric. The compartment that matters is visceral. The interventions that reduce visceral fat are not always the same as the ones that reduce total weight, which is why some people lose weight without metabolic improvement and others improve metabolically without losing much weight.
How to estimate it without a DEXA
Direct measurement requires imaging — DEXA, MRI, or CT scan. These are accurate but not universally accessible. Several proxy measures correlate reasonably well with VAT and are usable at home or with minimal equipment.
Waist-to-height ratio is the most useful single proxy. Measure the waist at the level of the umbilicus, divide by height in the same units. A ratio above 0.55 in adults indicates likely elevated VAT. Above 0.6 indicates a strong likelihood of metabolic risk regardless of total body weight. This metric outperforms BMI in predicting cardiovascular outcomes in large cohort studies.
Waist-to-hip ratio captures distribution. Above 0.90 in men or 0.85 in women indicates an android (apple-shape) distribution typically associated with higher VAT. This is rougher than waist-to-height but useful as a directional indicator.
What actually moves visceral fat
The interventions that preferentially reduce VAT are largely the same interventions that improve metabolic health more broadly — which is not coincidence. Visceral fat is responsive in ways that subcutaneous fat often isn't.
- Total caloric deficit, achieved gradually. VAT mobilises preferentially in early weight loss. A modest sustained deficit (300-500 calories below maintenance) produces disproportionately large reductions in visceral fat compared to total weight lost in the first 8-12 weeks.
- Resistance training, independent of weight loss. Multiple trials show resistance training produces visceral fat reduction even when total body weight is stable, via mechanisms including improved insulin sensitivity and increased lean mass.
- Reduction in liquid sugar intake. Fructose from sweetened beverages is metabolised primarily in the liver and contributes disproportionately to visceral and hepatic fat accumulation relative to its caloric contribution.
- Sleep duration consistently above 7 hours. Chronic sleep restriction below 6 hours nightly is independently associated with VAT accumulation, mediated through cortisol elevation and altered appetite regulation.
- Alcohol moderation. Ethanol is preferentially partitioned toward hepatic and visceral fat storage. Even moderate regular intake (above 7 drinks per week) shows measurable VAT accumulation in longitudinal studies.
- Aerobic exercise, particularly higher-intensity intervals. Total weekly aerobic volume matters less than total energy expenditure plus intensity for VAT reduction; both moderate-intensity steady-state and high-intensity interval training are effective.
The reframe
Body composition is not a single number. The lean-looking person with high VAT and the visibly heavier person with low VAT are at different metabolic risk — and the standard metrics most people track don't reliably distinguish between them.
If the goal is metabolic health rather than appearance, the question changes. Not "how much total fat do I carry?" but "where is the fat I carry stored?" The interventions that answer the second question well usually answer the first one too, with the additional benefit of producing measurable improvements in markers that actually predict whether the next two decades go well.