Obesity in the elderly: Should we be using new diagnostic criteria?

The proportion of elderly people is growing in almost every country. This conveys a higher rate of diseases associated with age among which nutrition-related disorders are many. Protein-energy malnutrition is highly prevalent (about 50%) in institutions and acute care units. Obesity is another issue that is raising increasing attention. The proportion of persons “diagnosed as obese” is higher in the elderly than in their young pairs. For example, in France, the latest survey conducted in 2006 (1) shows than the mean prevalence of obesity is 12.6, but more than 15% in those aged 65 years and above. Since obesity is associated with co-morbidities, it follows that at least one out of 5 of the patients in our wards are likely to be obese.

Obesity is diagnosed as a BMI (Weight/Height²) higher than 30 kg/m² (2). This is the criterion that is used to estimate the prevalence irrespective of age. However, although it is true that weight tends to increase with age (3), the increase in BMI is also associated with a reduced height. It is estimated that height decreases by 8 cm in women and by 5 cm in men between 20 and 85 years of age (4, 5). Therefore BMI naturally increases with age and mean values at 65 years are around 26 kg/m², higher than 25 and in the group of patients classified overweight. Each cm loss in height increases BMI between 0.2 and 0.3 units (for heights between 150 and 180 cms). Despite this mathematical artifact, the increased prevalence between 2003 and 2006, on the same French population (1) suggests that there is indeed an growing number of people with a high BMI as age increases.

Body composition changes with age: fat mass increases especially in the central area of the body and is said to double between 20 and 85 years of age (6). Fat-free mass decreases by 15% over the same age range and more than 2/3 of the decline is made of muscle (6, 7). Muscle tissue is infiltrated with triglycerides (8). This means that the BMI-% fat relationship varies with age. Figures 1 and 2 illustrate this concept, with a higher fatness for a given BMI as age increases. Furthermore, since there is a decrease in lean mass, a high BMI in a young adult (high lean mass, high fat mass, 9) corresponds to a different body composition in an older person (low lean mass, high fat mass). Further to this basic difference in body composition are the concepts of 1) sarcopenic obesity, with an increased fatness and a very low muscle mass, which carries specific consequences (10), and 2) an increased deposition of fat in the abdomen which drives metabolic consequences.

Figure 1.

Relationship between BMI and fatness. Data are taken in men from studies performed on subjects with varying age and classified as <60, 60-75, >75 years (3, 14, 15, 16)

Figure 2.

Relationship between BMI and fatness in women (3) age <50 or >60 years

The BMI cut-offs to define obesity have been chosen because of an increased all cause mortality as BMI increases above 25 kg/m² (2). In the elderly, this relationship is not the same. In a cohort of 13,000 subjects from California, aged 73 at study entry and followed up for 23 years, the adjusted relative risk of death is 1.35 before 70, 1.33 between 70 and 75, 1.12 above 75 and 1.22 above 80 years (11). Zamboni et al have reviewed (4) carefully this relationship. It looks as if the BMI-death relationship is weak when all cause mortality is considered, but that it is strong for cardiovascular death. It is also valid that the length of time exposed to excess weight is a key determinant in mortality. It appears that weight gain can be associated with increased mortality, but that involuntary weight loss also. Many confounding factors are listed among which smoking, metabolic co-morbidities and “reverse causality” (12). Therefore, while there probably remains a BMI-mortality relationship it is a difficult one to define obesity on.

Obesity should be defined as the amount of excess fat storage associated with elevated health risk. Since many health risks, including osteo-arthritis (4), loss of autonomy (13), and decreased quality of life are associated with increased fatness in the elderly, it may be more appropriate to define obesity on fatness than on BMI.

Financial disclosure: None of the authors had any financial interest or support for this paper.