Elsevier

Clinical Nutrition

Volume 29, Issue 5, October 2010, Pages 560-566
Clinical Nutrition

Original Article
The relationship between BMI and percent body fat, measured by bioelectrical impedance, in a large adult sample is curvilinear and influenced by age and sex

https://doi.org/10.1016/j.clnu.2009.12.011Get rights and content

Summary

Objective

The study aimed to establish the effects of age, gender and age–gender interactions on BMI–% fat relationships over a wide range of BMI and age. It also aimed to examine controversies regarding linear or curvilinear BMI–% fat relationships.

Methods

Body composition was measured using validated bio-impedance equipment (Bodystat) in a large self-selected sample of 23,627 UK adults aged 18–99 (99% ≤70) years, of which 11,582 were males with a mean BMI of 26.3 ± 4.7 (sd) kg/m2, and 12,044 females, with a mean BMI of 25.7 ± 5.1 kg/m2. Multiple regression analysis was used.

Results

BMI progressively increased with age in women and plateaued between 40 and 70 years in men. At a fixed BMI, body fat mass increased with age (1.9 kg/decade), as did % fat (1.1–1.4% per decade). The relationship between BMI and % fat was found to be curvilinear (quadratic) rather than linear, with a weaker association at lower BMI. There was also a small but significant age–gender interaction.

Conclusion

The association between BMI and % body fat is not strong, particularly in the desirable BMI range, is curvilinear rather than linear, and is affected by age.

Introduction

Many studies have examined the variation in body composition associated with age and gender.1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 These have generally used only modest numbers of subjects, as the measurements (with the exception of body mass index (BMI) and bioelectrical impedance) are often time-consuming. The usual pattern in most populations studied is that body fat and percent fat are greater in women than men, and increase until the age of 60 years or more.

The most commonly used indicator of % body fat is BMI (e.g. see validation studies against reference techniques9, 10, 13, 14, 15), although it is well known that it has an imperfect association. Muscle mass can vary considerably between individuals of the same height, and it contributes substantially to the variability in BMI, especially in leaner individuals. Many studies have drawn attention to the BMI–% body fat association, often in particular groups of subjects. Recent examples include athletes,16 military personnel,17 young adults,18 and the severely obese.19 There are also large population studies that provide reference values of body composition based on bioelectrical impedance analysis e.g.20, 21 However, in the general population there is uncertainty and some controversy as to whether the relationship between BMI and % body fat is linear,10 or curvilinear.9 There is further uncertainty about the magnitude of the increase in % fat with age after controlling for BMI,9, 11 and even more uncertainty as to whether the age-dependent change in % fat relationship at a given BMI is affected by gender. We wish to add to this body of evidence by analysing a collection of data where body composition is measured directly in a large sample.

We have accumulated a large set of data on body composition in UK adults using bio-impedance measurement equipment (Bodystat Ltd, Douglas, Isle of Man, UK). Validation studies in which % body fat is the outcome variable, have been undertaken in lean and obese Caucasian subjects without disease using Bodystat bielectical impedance equipment and a range of reference techniques, such as the 4 compartment model,22 3 compartment model,22, 23, 24 total body water,22, 23 hydrodensitormetry,22, 23 dual energy X-ray absorptiometry (DXA),22, 25, 26 and air displacement plethysmography.27 Bodystat instruments have also been shown to have good test–retest reliability with respect to fat and fat-free mass and in the raw measurements used to estimate body composition.25, 28 We now use this new large data set to examine the significance of gender, age–gender, and age–BMI interactions in predicting percent body fat. In doing so we also look at the extent to which changes with age are due to alterations in lean body mass and fat mass.

Our study has the strength of using a large sample of men and women over a wide range of BMI and age across the UK. A drawback is that subjects are largely self-selected, and so we are unable to reliably derive estimates of UK population averages. However, it is still reasonable to expect that the patterns of body composition variation with age and gender are representative of the UK population.

Section snippets

Subjects

Subject data were obtained from Bodystat® 1500 bio-impedance analyzers (Bodystat Ltd, Douglas, Isle of Man, UK) returned to the company for servicing during the period 2000–2006. Although subjects were not informed that the data might be used for research, they were completely anonymous – the equipment does not store information that would enable subjects to be identified. Subjects will generally have been healthy, as the devices were not used for medical purposes. Few individuals had a BMI of

Body fat percentage, fat mass, and lean mass

Fig. 1 shows mean % body fat as a function of age and gender. The pattern is clearly one of a fairly steady increase from age 20 to 70 in both genders. The increase amounts to 2.4% (se 0.06%) and 1.9% (se 0.05%) per decade for females and males respectively if a linear trend is fitted. The mean becomes more variable after the age of 60 years due to the smaller subject numbers.

Fig. 2 shows that the increase in % body fat with age is predominantly due to a steady increase in fat mass and a

Discussion

The changes in body composition with age have been widely studied, and so we cannot claim novelty for some of the results we have presented in this paper. The contribution we have made is to verify the patterns that other studies have shown by using a well-tested method on an exceptionally large sample of adults spanning a wide range of age and BMI. That the patterns in our data in the UK are similar to those found elsewhere gives us confidence that the aspects of the BMI–body composition we

Conflict of interest

ME and GWH have no conflict of interest. SM (Managing director of Bodystat Ltd) provided the data collected by the Bodystat bio-impedance machines to ME and GWH, both of whom analysed the data and drafted the paper.

Author agreement

All authors agreed on this version of the paper. SM collected the data and some references, and the other two authors drafted the paper.

Acknowledgements

We are grateful to the many organisations from whom the data were collected, and to all the individual subjects involved. This work was partly supported by the Scottish Government Rural and Environment Research and Analysis Directorate.

References (39)

  • U.G. Kyle et al.

    Increased length of hospital stay in underweight and overweight patients at hospital admission: a controlled population study

    Clin Nutr

    (2005)
  • J. Woo et al.

    Body composition in Chinese subjects: relationship with age and disease

    Arch Gerontol Geriatr

    (1998)
  • A.S. Jackson et al.

    The effect of sex, age and race on estimating percentage body fat from body mass index: the Heritage Family Study

    Int J Obes Relat Metab Disord

    (2002)
  • D. Gallagher et al.

    How useful is body mass index for comparison of body fatness across age, sex, and ethnic groups?

    Am J Epidemiol

    (1996)
  • P. Deurenberg et al.

    Body mass index as a measure of body fatness: age- and sex-specific prediction formulas

    Br J Nutr

    (1991)
  • U.G. Kyle et al.

    Age-related differences in fat-free mass, skeletal muscle, body cell mass and fat mass between 18 and 94 years

    Eur J Clin Nutr

    (2001)
  • P. Deurenberg et al.

    Body mass index as a measure of body fatness in the elderly

    Eur J Clin Nutr

    (1989)
  • A.S. Jackson et al.

    Generalized equations for predicting body density of women

    Med Sci Sports Exerc

    (1980)
  • D. Black et al.

    Obesity. A report of the Royal College of Physicians

    J Roy College Physicians London

    (1983)
  • Cited by (198)

    View all citing articles on Scopus
    View full text