FFMI and the natural limit debate: what the 1995 study actually found

Where "FFMI 25" comes from

The idea that a normalized fat-free mass index of 25 marks the edge of what's achievable without anabolic steroids traces back to a single 1995 study. Kouri, Pope and Katz measured fat-free mass index in 157 male athletes — 83 self-reported anabolic-androgenic steroid users and 74 non-users — and published the results in the Clinical Journal of Sport Medicine (PubMed). Among the non-users, normalized FFMI values extended up to a well-defined ceiling around 25.0; among the users, values commonly ran from 25 up into the 30s. That single dataset is the entire origin of "FFMI 25" as a natural-limit shorthand — it's cited constantly, but it's one study of a specific population, not a biological law.

What FFMI measures, and why it's normalized

Fat-free mass index divides lean (fat-free) body mass in kilograms by height in meters squared — the same structure as BMI, but for muscle rather than total weight. Raw FFMI is biased by height, because taller people carry more frame and bone mass independent of how muscular they are, so the field applies a height correction: normalized FFMI = FFMI + 6.1 × (1.8 − height in meters), which adjusts every result to what it would be at 1.80 m. This is the exact formula the FFMI calculator on this site uses, and it's the version the natural-limit research is built around — raw FFMI numbers aren't directly comparable to the 25 figure without it.

Worked example: normalization changes the story

Take a lifter at 165 cm, 58 kg, 22% body fat — 45.2 kg of lean mass. Raw FFMI: 45.2 ÷ 1.65² ≈ 16.6. That looks unremarkable next to the "25" headline number, but normalized to 1.80 m it becomes 16.6 + 6.1 × (1.8 − 1.65) ≈ 17.5 — still below average, but the correction matters more the further someone's height sits from 1.80 m. Now take a taller lifter at 190 cm, 100 kg, 10% body fat — 90 kg of lean mass. Raw FFMI: 90 ÷ 1.90² ≈ 24.9, close to the natural-limit figure already. Normalized: 24.9 + 6.1 × (1.8 − 1.90) ≈ 24.3 — the correction pulls it down slightly, because at 190 cm some of that raw FFMI is explained by height alone. Comparing the two lifters' raw numbers (16.6 vs. 24.9) overstates the real gap in muscularity; the normalized numbers (17.5 vs. 24.3) are the fairer comparison.

What the 25 boundary does and doesn't mean

The Kouri paper itself frames 25 as a descriptive upper bound observed in one sample, not a diagnostic cutoff for detecting steroid use in an individual. The non-user group's mean normalized FFMI was 21.8 (± 1.8) — most drug-free athletes in the study sat well below 25, and 25 represented the outer edge of that distribution, not the typical result. The paper also reports that a separate historical sample of 20 Mr. America winners from the pre-steroid era (1939–1959) averaged a normalized FFMI of 25.4 — meaning the very population most often cited as the natural ceiling includes a group of elite, drug-free physique champions whose average sits just above the commonly quoted number. That detail alone should temper how absolute "25" gets treated in casual use.

Where the metric breaks down

Three limitations are worth naming plainly. First, "non-user" status in the study was self-reported, not drug-tested — some misclassification is possible in either direction. Second, the sample was specifically competitive male athletes and bodybuilders, a population selected for high muscularity and genetic responsiveness to training; it says little about the typical range for the general population, and the formula hasn't been re-validated at this scale across broader groups, ages or women. Third, and most practically: FFMI is only as accurate as the body-fat percentage used to calculate lean mass. As covered in the body-fat estimation comparison on this site, field methods like tape measurements and bioelectrical impedance carry error ranges of roughly ±3–10%, and that error propagates directly into the FFMI result — a body-fat reading that's off by 5 percentage points can shift someone's normalized FFMI by a meaningful margin without their actual muscle mass changing at all.

A single study, treated as more than it is

None of this means FFMI is useless — it's a reasonable, height-adjusted way to track changes in lean mass over time in the same person, measured consistently. The issue is specifically the "25 = natural limit" framing, which compresses one 1995 study of 157 athletes into a rule that gets applied to individuals with very different genetics, training histories, limb proportions and body-fat measurement methods than anyone in the original sample. Treat a normalized FFMI near or above 25 as worth noting, not as evidence of anything about how a specific person achieved it.

Using it sensibly with the tools

The FFMI calculator on this site reports both raw and normalized FFMI plus a category, using the same formula described above; the lean body mass calculator is the input worth getting right first, since an inaccurate body-fat percentage is the most common source of a misleading FFMI result. Track the normalized number over months as a consistency check on real muscle gain, rather than treating a single reading — yours or anyone else's — as proof of anything.

---

This article explains a published research study and is general information, not an assessment of any individual's training status, health or substance use. It is not medical advice.