Now is the time of year when many people most dread stepping onto the bathroom scale. But, while all of that holiday feasting may have led to greater girth, how should we understand what higher or lower body weight means?
The body mass index (BMI) has long been used in epidemiology, medicine, and nutritional sciences. But its value is increasingly being questioned, especially in obesity research, where body-composition measurements are proving to be far more relevant.
The BMI is calculated by dividing a person’s body weight in kilograms by the square of his height in meters (kg/m2). First named after the Belgian astronomer, mathematician, and statistician Adolphe Quetelet, who demonstrated in 1835 how adult weight normally increases in proportion to height squared, the index provides a measure of body weight, independent of stature, allowing us to compare the weight of short and tall people.
In 1972, an American scientist, Ancel Keys, working in the area of human nutrition, public health, and epidemiology, renamed it the BMI after finding that its values also correlated with body fat mass as derived from skin-fold or body-density measurements.
However, recent findings have cast doubt on the BMI’s value. For example, although the BMI is associated with fat mass in obese people, there is little or no association with that of normal or underweight individuals. At any given BMI score, fat mass varies widely, and other variables, such as gender or age, further distort findings, with greater increases in fat mass per BMI unit seen in women and in the elderly.
Although the BMI is an unscientific way to characterise a person’s nutritional status, it is nonetheless used as a measure of total body fat mass in medical practice and epidemiological studies, not least because it is easy to calculate and document regularly in personal health records. Typically, doctors use it to categorise patients as “underweight” (a BMI value below 18.5), “normal weight” (18.5-25), “overweight” (25-30), or “obese” (above 30).
This categorization is determined by data drawn from the general population and based on the assumption that there is a higher risk of metabolic or cardiovascular disease found at the high end (and sometimes the low end) of the BMI scale. Although recent studies suggest that high BMI scores do not necessarily increase risk of death, doctors find the BMI categories a useful basis for prevention and treatment.
Such usage, however, has serious limitations. Because the index is calculated from two biological measures (weight and height), the resulting score has no biological meaning in itself. Thus, studies of the genetics of obesity that are based on a relationship between certain genetic markers and the BMI are also meaningless.
In fact, the use of the BMI is likely to obscure our understanding of genetic effects on body weight. This is because body weight is the sum of our organs and body tissues; and each organ or body part has its own regulatory (and thus partly genetic) basis. We can understand more about organ and tissue masses, for example, or fatty infiltrations in individual organs like the liver and pancreas, by focusing on the nature of each body component, rather than by relying on an overall body score.
Indeed, the regulation of body weight overall is determined by the sum of the specific regulatory outcomes affecting individual body components. Because individual body components are inter-related, control of body weight seems to occur in the relationships between tissues and organs, rather than within individual components.
Adjusting weight by height squared is also likely to yield a different outcome for different organs. Although many body components do scale to height by a power close to two, others do not. The brain, bone, and mineral mass scale to height by a power greater than two; while fat mass scales to height by a power of 1.8-2.6, depending on the population being studied and the means of measurement used.
In short, body weight and body fat mass may not scale to height by the same power. This power may vary by population; and short and tall subjects with equal BMI scores, from the same population, may have a different body composition.
Further evidence supporting the use of body-composition analysis, rather than the BMI, comes from the metabolic heterogeneity observed within each BMI category. For example, a subgroup of normal-weight subjects may have low subcutaneous fat mass but high visceral fat mass (“thin on the outside and fat on the inside”). Despite their normal BMI, metabolically obese individuals – who may comprise as much as 24 per cent of a normal-weight population – may be insulin-resistant and suffer increased cardio-metabolic risks. Likewise, around half of overweight subjects, and 15-45 per cent of obese subjects, appear to have a favorable metabolic profile (that is, no metabolic complications, inflammation, dyslipidaemia, or hypertension).
At a BMI above 30, these subjects are deemed “metabolically healthy obese.” Like obese patients suffering from cardio-metabolic risks (the “metabolically abnormal obese”), they have high visceral and subcutaneous fat mass, though metabolically healthy obese patients sometimes have excessive body fat but lower fat infiltration in the liver and in skeletal muscle.
But the differences between the metabolically healthy obese and the metabolically abnormal obese are inconsistent. And, though age, sex, ethnicity, waist circumference, physical activity, smoking, and alcohol consumption all correlate with metabolic phenotypes, the BMI is unable to distinguish between the two groups.
It would be unfair to say that the BMI has no clinical value; it can be a useful indicator of nutritional health in patients, and help doctors make day-to-day decisions about whom they should treat. But one must turn to body-composition measurements for any serious epidemiological and etiological studies on obesity.
Manfred J. Müller is Professor of Physiology at Christian Albrechts University in Kiel, Germany.
Copyright: Project Syndicate, 2014.