Relative fat mass (RFM)

Clinical background

BMI is the most widely used clinical method for classifying overweight and obesity, but it does not distinguish fat mass from fat-free mass. This leads to systematic misclassification: a muscular patient may be given too high a BMI, and a normal-weight person with a high fat percentage may be missed. In the derivation study, BMI had a sensitivity of only 21.6% for obesity defined by DXA-measured body fat in women, and 81.9% in men [1]. Relative fat mass (RFM) was developed to provide a simple alternative that requires only a tape measure and that estimates body fat percentage better than BMI.

Calculating relative fat mass

RFM is calculated as:

$$\text{RFM} = 64 - 20 \times \frac{\text{height}}{\text{waist circumference}} + 12 \times \text{sex}$$

where sex is 0 for male and 1 for female, and height and waist circumference are given in the same unit (the ratio is dimensionless). The formula can also be written separately: for men $64 - 20 \times (\text{height}/\text{waist circumference})$ and for women $76 - 20 \times (\text{height}/\text{waist circumference})$.

RFM was derived from NHANES 1999–2004 (n = 12,581 adults ≥20 years) with DXA as the reference method for body fat. Among 365 anthropometric indices generated, the height/waist circumference ratio was chosen because it was the simplest index with the best performance, particularly in men. Validation was carried out in a separate cohort, NHANES 2005–2006 (n = 3,456) [1]. The study population consisted of American adults of Mexican, European and African ethnicity.

Interpretation in practice

RFM gives an estimate of total body fat percentage and should be interpreted as such, not as a measure of visceral adipose tissue. Since the calculator does not define its own interpretation bands, reference is made to the cut-offs published by the developers themselves, based on the relationship between RFM and all-cause mortality in NHANES 1999–2014 (n = 31,008, median follow-up 8.3 years) [2]:

These cut-offs were confirmed in an independent cohort (NHANES III, n = 12,650, median follow-up 23.3 years) [2]. A prospective study of 26,754 adults in NHANES 1999–2010 found that mortality risk rose markedly when RFM exceeded 30% in men and 45% in women, with a U-shaped relationship in men but not in women [3].

Clinically, this means that a man with an RFM below 30% and a woman with an RFM below 40% probably do not have obesity as defined by body fat, whatever the BMI shows. If the RFM is above the cut-off, the patient should be assessed for metabolic risk and lifestyle intervention considered, but RFM does not replace a complete cardiometabolic risk assessment.

Validation and performance

In the validation cohort of the derivation study, RFM was superior to BMI in predicting DXA-measured body fat percentage, with R² = 0.75 in men (versus 0.61 for BMI) and R² = 0.69 in women (versus 0.65 for BMI) [1]. RFM had lower bias than BMI in women (0.9% versus −10.9%) and a similarly low bias in men. Total misclassification of obesity was lower with RFM than with BMI in both women (12.7% versus 56.5%) and men (9.4% versus 13.0%) [1].

A systematic review and meta-analysis from 2025, which included 10 studies of RFM (n = 20,230 for correlation analysis, n = 33,687 for mean difference), found a moderate pooled correlation between RFM and reference-measured body fat of r = 0.77 (95% CI 0.72–0.81, I² = 97%) [4]. The mean difference between RFM-estimated and observed body fat was −0.04 percentage points (95% CI −0.93 to 0.85), that is, without systematic over- or underestimation in the pooled analysis [4]. In adults, however, RFM tended to overestimate body fat by an average of 0.76 percentage points [4].

External validation has been performed in several populations. In a Mexican cohort (n = 61, aged 20–37 years), RFM was a better predictor of DXA-measured fat percentage than BMI (R² = 0.84), but the intercept deviated from zero when RFM was compared with air displacement plethysmography, bioimpedance and a four-compartment model, suggesting method-dependent bias [5]. In a Chilean cohort (n = 270), the AUC for RFM and BMI did not differ significantly for identifying raised body fat (men: 0.970 versus 0.959; women: 0.946 versus 0.942), and Bland–Altman analysis showed more pronounced bias in men than in women [6].

In a prospective study of 46,535 adults in NHANES 1999–2018 (median follow-up 9.7 years), RFM was more strongly associated with diabetes-related mortality than either BMI or waist circumference, while the association with cardiac mortality and all-cause mortality was similar across the three measures [7]. The results were reproduced in NHANES III (n = 14,448) [7].

Limitations

RFM was derived in an American population and has been validated mainly in North and South American cohorts. The review found high heterogeneity (I² > 75%) in all meta-analyses, reflecting differences between populations, reference methods and measurement technique [4].

Performance declines with age. In the derivation study there was a significant interaction between age and RFM in women, and all models performed less well in older individuals, probably because of age-related changes in body composition with reduced fat-free mass and a shift in fat distribution [1]. RFM has not been validated sufficiently in children and adolescents, and a paediatric version (RFMp) has been proposed but is not established.

Waist circumference measurement technique is decisive. NHANES uses a specific method in which the waist is measured at the upper lateral border of the right ilium. If the technique deviates from this, the RFM value shifts. This is the single most important source of error in clinical use.

RFM estimates total body fat percentage, not visceral fat. Nor does it distinguish between subcutaneous and intra-abdominal adipose tissue, and it gives no information about fat-free mass or muscle mass. A patient with sarcopenia and a normal RFM may still have an unfavourable body composition.

References

1. Woolcott OO, Bergman RN. Relative fat mass (RFM) as a new estimator of whole-body fat percentage: A cross-sectional study in American adult individuals. Sci Rep. 2018;8(1):10980. PMID: 30030479
2. Woolcott OO, Bergman RN. Defining cutoffs to diagnose obesity using the relative fat mass (RFM): Association with mortality in NHANES 1999–2014. Int J Obes (Lond). 2020;44(6):1301–1310. PMID: 31911664
3. Wang J, Guan J, Huang L et al. Sex differences in the associations between relative fat mass and all-cause and cardiovascular mortality: A population-based prospective cohort study. Nutr Metab Cardiovasc Dis. 2024;34(3):738–754. PMID: 38161128
4. Palumbo AM, Jacob CM, Khademioore S et al. Validity of non-traditional measures of obesity compared to total body fat across the life course: A systematic review and meta-analysis. Obes Rev. 2025;26(6):e13894. PMID: 39861925
5. Guzmán-León AE, Velarde AG, Vidal-Salas M et al. External validation of the relative fat mass (RFM) index in adults from north-west Mexico using different reference methods. PLoS One. 2019;14(12):e0226767. PMID: 31891616
6. Aguirre C, Tumani MF, Carrasco F et al. Relative fat mass as an estimator of body fat percentage in Chilean adults. Eur J Clin Nutr. 2024;78(9):782–787. PMID: 38942896
7. Woolcott OO, Samarasundera E, Heath AK. Association of relative fat mass (RFM) index with diabetes-related mortality and heart disease mortality. Sci Rep. 2024;14(1):30823. PMID: 39730510