Beyond the Scale and Tape Measure: Measurement of Obesity

What is obesity? “Obesity” has recently received unprecedented attention from science, healthcare, education, government, and community organizations. However, although the definition, excess body fat, seems straightforward, actual measurement brings to light many questions. How much body fat is too much? Where in the body is excess adiposity a problem? Do amount and location have different implications for different people? What are valid measures of the amount and location? For indirect measures, what are appropriate cutpoints to define excess?

Assessing Adiposity, a Scientific Statement from the American Heart Association, provides a comprehensive, evidence-based review of this subject. The information is practical guidance for not only clinical researchers who seek to identify precise measurement for their subjects, but also clinicians who care for patients for whom excess weight is a clinical problem (and isn’t that most clinicians?) and all who read the scientific literature on human adiposity.

Obesity is a heterogeneous condition. The variation has been a focus of study for over 50 years.¹ The existence of individuals who are obese but metabolically healthy proves the complexity of the relationship between body fat and metabolic health. As a result, scientists have looked for patterns in this heterogeneity and examined regional distribution. The development of imaging techniques has allowed measurement of body fat topography and therefore the ability to distinguish between abdominal subcutaneous fat tissue and the visceral fat that can collect in the abdominal cavity. From this finding came the observation that visceral adipose tissue is associated with greater cardiometabolic risk. However, this relationship is complex; vulnerability to visceral fat deposition varies with sex, age, and race; it is more common among women, among the elderly, among Asians compared with Caucasians and among Caucasians compared with African Americans.^2-4 More precise imaging techniques have allowed study of ectopic fat storage, in particular in the liver, and its relationship with visceral fat. Liver fat is highly predictive of cardiometabolic abnormalities, and studies to understand the pathophysiology of this relationship are a high priority.⁵

Precision or practicality: the review of total body adiposity measurement acknowledges the tradeoff. Body mass index (BMI) is simple and reproducible and works very well for epidemiologic studies but in individuals cannot distinguish high lean body mass from fat or capture any information about distribution. Despite the limitations in sensitivity and specificity, BMI remains the most practical method in clinical practice. Waist circumference is also simple and inexpensive and can predict visceral adiposity. The preferred site of assessment and cutpoints are laid out. Although hip circumference, neck circumference, and various ratios have all been evaluated, for clinical use waist circumference remains the one measure with a track record of utility in adults.⁶ CT and MRI are precise but cumbersome methods of adipose tissue distribution and body composition reserved for research purposes. For the non-expert, the statement explains how these imaging methods distinguish fat from lean body, and the standard location of the imaging sequences.

Methods to measure not only body fat but other components of body composition, such as muscle and bone, include “low-tech” approaches, like skinfold thickness and mid upper arm and mid thigh circumferences. Although these methods may appear easy, accuracy and reliability are operator-dependent, and even among experienced assessors, the methods are compromised by obesity and by loose connective tissue in the elderly. Also, their use requires complex prediction equations to extrapolate percent body fat from the 3 to 7 measures. Near infra-red interactance uses the principle that optical densities are inversely related to percent body fat. A light probe placed on the skin at standard sites, such as thigh or abdomen, provides a quick, non-invasive and potentially inexpensive measure, but studies to date have not demonstrated good validity.⁷ Similarly, bioelectrical impedance is safe, quick, and low cost, but validity varies with not only sex, age and race-ethnicity, but other factors that may change from day-to-day such as environment and phase of menstrual cycle.⁸ Hydrostatic weighing, historically a gold standard, is not only challenging to do but can lead to inaccurate results because of inherent limitations in accounting for air in the GI tract and lungs as well as different assumptions about fat and fat-free density.⁹ Air displacement plethysmography measures as accurately as hydrostatic weighing and is quick and safe but does not assess fat distribution. Dual energy X-ray absorptiometry (DEXA) requires expensive equipment, but it is an excellent research tool because it provides information on fat distribution as well as body composition and it is safe and quick.¹⁰

The development of a high quality assessment of ectopic fat will fulfill a huge need in the field of non-alcoholic steatohepatitis, where clinicians and scientists must use liver biopsy for valid measure because of the near uselessness of non-invasive screens such as liver enzymes or ultrasound. Localized proton magnetic resonance spectroscopy is emerging as a non-invasive method of quantifying liver fat.¹¹

Serial measurement adds more challenges. The benefits of weight loss may seem obvious; however, loss of kilograms inevitably includes loss of fat-free mass, which can dampen the benefit of fat loss. Measuring loss of kilograms provides no information about change in body composition and fat distribution. Studies on serial waist circumference change have shown conflicting evidence of validity.¹² However, air displacement plethysmography, DEXA, and BIA all appear to accurately capture change in body composition. Further work on the strengths and limitations of longitudinal use of each measurement method is ongoing.

An ongoing focus of studies is the variation in validity and meaning of each method among specific subpopulations defined by race-ethnicity, sex, and age, Although children are not addressed specifically in this statement, testing for validation, reliability, and implications of these methods in infants through adolescents is also required.

The high prevalence of obesity in the United States and the urgent need to uncover effective interventions through both physiologic and behavioral studies has made this problem a high priority for research. Improvement in measurement in this field is essential to that work.