There are many reasons to estimate physiological processes related to obesity and diabetes in epidemiologic and genetic studies. Such studies, which often have large numbers of participants, are often designed to examine the importance of explicit risk factors in the pathogenesis of disease. In genetic studies, estimation of specific processes (e.g., insulin resistance, β-cell function, immune system function) allow for the performance of quantitative trait analysis. The traditional approach in genetics has been to compare those with to those without disease. But, the realization that complex conditions such as obesity, and diseases such as diabetes often take years to emerge leads us to identify certain risk factors which may predict disease years before emergence, and examine genetic loci for the risk factors. It is a major challenge to assess function in large numbers of subjects. If there is a single clinical measurement which reflects underlying biology, the issue is straightforward. The progress of proteonomics and metabolomics holds promise for finding simple correlates of underlying function. However, simplicity is rarely the case. The problem of finding acceptable surrogates has been a particularly difficult one in the study of metabolic conditions. Obesity is a risk factor for a variety of diseases. It is well documented that visceral fat deposition is more of a risk than that of subcutaneous fat deposition, and there are gender and ethnic differences in relative fat deposits. Thus, there is a need to assess total stored fat as well as relative depots. Computed axial tomography scans and magnetic resonance imaging to measure fat deposition have been major advances. But, it is often not practical to use these expensive methods (possibly involving radiation), so that in many studies BMI is used for total fat and waist-to-hip ratio, or waist circumference is used as an index of visceral fat deposition. Problems are even greater for diabetes risk. How does one assess major risk factors: insulin resistance, β-cell function, and/or glucose effectiveness in a large study without doing invasive tests? The emergence of surrogates such as homeostasis model assessment (HOMA) indices (1), quantitative insulin sensitivity check index (QUICKI) (2), and a bevy of assessments from the oral glucose-tolerance test (OGTT) have emerged (3,4,5). These indices are seductive, because they are very easy to calculate. Some of them are virtually identical; QUICKI is a mathematical transformation of the HOMA insulin resistance index and adds no additional information. HOMA insulin resistance index will provide little extra information beyond fasting insulin in a nondiabetic population with relatively similar glucose values. OGTT-based indices can be more useful, but remain to be rigorously validated. Nevertheless, there can be little argument that surrogate measures are important for large epidemiologic and genetic studies, and there is a critical need to continue to evaluate them for accuracy and sensitivity. A problem arises when surrogate measurements are used in the clinical setting. BMI itself must be used with great care. It is important to remember that BMI is related to body weight and the height squared. Thus, comparing BMI values among individuals of different stature, and those with widely varying musculature may not be appropriate. Short individuals with large thorax will have greater BMI than individuals with narrower bone structure, independent of adiposity. Several publications have documented the rather weak relationship between BMI and adiposity (6,7). There are clear ethnic differences in the relationship between BMI and body fat (8). Yet, it is not clear that practitioners will take such differences into account. Similarly, using HOMA or its essentially equivalent QUICKI in individual nondiabetic patients may not reflect degree of insulin resistance much better than is revealed by the fasting insulin value. It is even more difficult to assess β-cell function in individual patients. If fasting insulin (or HOMA) reflects insulin resistance, can it also reflect β-cell function? In fact, using the ratio of C-peptide to glucose is likely a better surrogate for β-cell function, as plasma insulin depends upon first pass and peripheral insulin clearance, both of which are changed by obesity and other insulin-resistant conditions. The cautionary note based upon the preceding considerations is that “high-faluting” numbers such as BMI and insulin resistance indices may not be better than body weight and fasting insulin to assess individual subjects. In any event, they must be used with great care when comparing different individuals of varying gender, ethnicity, and body habitus.
Search from 250 M+ research papersSearch from 250 M+ research papers
Feb 1, 2010
P Dongiovanni + 16
Open Access