What is the ideal body weight?

This paper attempts to answer five questions. (1) How should we best measure overweight? Although life insurance tables and relative weights have been used, I propose that body mass index (BMI; wt/ht2) is the preferred method. It is currently used in most epidemiologic studies and can be used in clinical evaluation of individual patients. (2) What is a healthy body weight? A healthy body weight can be defined from normative population data as well as from mortality experience. The National Center for Health Statistics has used the normative approach by defining overweight as a BMI above the 85th percentile of the weight distribution curve. Super overweight is defined as weights above the 95th percentile. There are three problems with this approach. First, the normative weight changes as the population weight changes, so standards differ with each survey. Second, this approach automatically defines 15% of the population as overweight. Finally, and of most concern, is the implication that a normal weight is a healthy weight. The alternative approach is to use mortality data. The life insurance companies have done this, using this data as the basis for the Metropolitan Life Insurance Table. This table uses three weight ranges based on frame size. However, frame size was never measured, and it appears that they were assigning the smallest quartile to small frame, the middle two quartiles to average frame, and the heaviest quartile to the large frame. The problem with this approach is that the same weight range is used for all ages and that the standard differs for men and women. A third approach is to use the mortality experience from large epidemiologic studies. Here the lowest death rates can be defined as the best and a range around this can be considered "healthy." Using this risk approach, BMI units separated by five BMI unit intervals have been defined, beginning with the healthy range of BMI between 20 and 25/27 kg/m2. (3) Should we also measure body fat? I will conclude that, except for research purposes, there is no clinical value in measuring body fat. (4) Should we measure body fat distribution? Because body fat distribution is related to health risks, a measure of fat distribution is important. (5) What are some other measures that may be needed to establish whether a given weight is healthy? The first of these is the amount of weight gain since age of 18 to 20 years. The second is the presence of comorbidities: The fasting insulin level can be used to assess the degree of insulin resistance; blood pressure measurement will determine the presence of hypertension: measuring triglyceride and high-density lipoprotein cholesterol both reflect hyperinsulinemia and insulin resistance. Finally, the presence of sleep apnea is an important variable that must be noted. With this information, it is then possible to use an algorithm to decide whether an individual is overweight and, if so, whether this is a risky form of overweight.