What is the "J-curve" in BMI and mortality research?

It describes the shape researchers commonly find when plotting BMI against average mortality risk: risk is higher at both the low and high ends of the BMI range, with the lowest average risk somewhere in the middle — not at the lowest possible BMI. The curve looks roughly like the letter J.

Does this mean being underweight is riskier than being overweight?

Research generally finds elevated mortality risk at both ends of the BMI spectrum, and the exact shape and location of lowest risk varies across studies and populations. It does not support the idea that "lower is always safer" — very low BMI carries its own documented risks.

Why might being underweight carry health risk?

Very low BMI can reflect inadequate nutrition, an underlying illness, or insufficient energy reserves, all of which are independently associated with worse health outcomes — a different mechanism from the risks associated with a high BMI.

What is the practical takeaway from the J-curve?

That the relationship between BMI and mortality risk is not a simple straight line where lower is always better. A BMI within the standard healthy range is generally associated with the lowest average population risk in most research, but individual health context matters more than chasing the lowest possible number.

BMI and the Mortality J-Curve

Research plotting BMI against average mortality risk commonly finds a J-shaped curve, not a straight line — risk is elevated at both the low and high ends, with the lowest average risk somewhere in the middle.

What the curve shows

Multiple large population studies have found that average mortality risk rises at both very low and very high BMI, with the lowest risk typically found somewhere within or near the standard healthy-to-mildly-overweight range, depending on the specific study and population. This is why "lower BMI is always safer" is not what the underlying research actually supports.

Why the low end carries risk too

A very low BMI can reflect inadequate nutrition, an undiagnosed underlying illness, or insufficient energy reserves — all independently linked to worse health outcomes through mechanisms unrelated to the risks associated with a high BMI. Some of the elevated risk at the low end of BMI research also reflects pre-existing illness that caused weight loss, rather than low weight itself causing the risk — an important nuance researchers continue to account for.

Reading this responsibly

The J-curve is population-level research, not an individual prescription. It does not mean a specific BMI number guarantees the lowest possible risk for any one person — genetics, fitness, existing conditions, and many other factors matter alongside BMI. A healthcare provider is far better positioned to interpret an individual's risk than a population curve alone.

Common questions

What is the "J-curve" in BMI and mortality research?: It describes the shape researchers commonly find when plotting BMI against average mortality risk: risk is higher at both the low and high ends of the BMI range, with the lowest average risk somewhere in the middle — not at the lowest possible BMI. The curve looks roughly like the letter J.
Does this mean being underweight is riskier than being overweight?: Research generally finds elevated mortality risk at both ends of the BMI spectrum, and the exact shape and location of lowest risk varies across studies and populations. It does not support the idea that "lower is always safer" — very low BMI carries its own documented risks.
Why might being underweight carry health risk?: Very low BMI can reflect inadequate nutrition, an underlying illness, or insufficient energy reserves, all of which are independently associated with worse health outcomes — a different mechanism from the risks associated with a high BMI.
What is the practical takeaway from the J-curve?: That the relationship between BMI and mortality risk is not a simple straight line where lower is always better. A BMI within the standard healthy range is generally associated with the lowest average population risk in most research, but individual health context matters more than chasing the lowest possible number.