The disposition index, abbreviated DI, is an important concept in both theoretical and clinical diabetology. Serving as a biomarker for the function of the insulin-glucose feedback control system, it can be determined using various methods and provides an estimate of the loop gain of the homeostatic system. Generally, it is calculated as the product of pancreatic beta-cell function times insulin sensitivity.
Under physiological conditions, insulin sensitivity and beta cell function adapt to each other over a prolonged period[1]. This means that in cases of ‘healthy’ obesity, i.e. without diabetes mellitus, beta cell mass slowly increases as insulin sensitivity decreases (i.e. insulin resistance increases). Conversely, beta cell function decreases in conditions of high insulin sensitivity (e.g., in prolonged starvation). This long-term adaptation of glandular mass to demand is a general principle in endocrinology, which is referred to as dynamical compensation[2].

The relationship between insulin sensitivity and beta-cell function is hyperbolic in a ‘metabolically healthy’ organism. Via dynamic compensation, beta-cell mass increases with progressive insulin resistance, whereas the disposition index, which is the product of insulin sensitivity and beta-cell function (green line), remains constant (transition from A to B). If this does not happen or is insufficient (transition from A to C), the result is prediabetes or type 2 diabetes mellitus.
The disposition index, which is the product of insulin sensitivity and beta-cell function, remains constant in healthy organisms despite variations in body composition and insulin resistance. If, however, the pancreatic beta cells are no longer able to compensate for increasing insulin resistance by increased cell mass, the disposition index decreases and glucose homeostasis is impaired. This is the case in advanced metabolic syndrome. Depending on the extent of the reduction in the disposition index, prediabetes or type 2 diabetes mellitus may develop.
For determining the disposition index, data on insulin sensitivity and beta-cell function have to be obtained. This can be done in various ways:
If based on a glucose clamp investigation, the disposition index is calculated as the product of the insulin response curve (AUCΔInsulin) and the insulin sensitivity index (ISIClamp) with
ISIClamp is the ratio of the mean glucose infusion rate and the mean insulin concentration during the clamp:

Using the frequently sampled intravenous glucose tolerance test (fsIGT), the time series of insulin and glucose concentration are to be fitted to the minimal model of glucose homeostasis by Bergman and Cobelli. Then the disposition index can be calculated with

from the first phase of the insulin response (φ1) and the insulin sensitivity index (S1)[43].
If the frequently sampled oral glucose tolerance test (fsOGT) is used, it is to be coupled with the approach of Matsuda and DeFronzo. Then, the disposition index can be calculated with

from the insulinogenic index (IGI) and the insulin sensitivity index (ISIcomposite)[54].
This method delivers a static disposition index that requires fasting concentrations of insulin and glucose only. It is calculated with

as the product of the secretory capacities of pancreatic beta cells (SPINA-GBeta) and the reconstructed insulin receptor gain (SPINA-GR)[65][76].
The approaches presented provide similar but slightly different information. Even though the results of the four methods correlate significantly with each other[76], several biological factors may cause certain deviations. These are rooted, for example, in the different mechanisms that determine static and dynamic insulin secretion patterns. In a comparative study, the discriminatory power of the static disposition index for the diagnosis of diabetes mellitus was higher than that of the oGTT-based disposition index according to Matsuda and DeFronzo[76].
The disposition index is reduced in diabetes mellitus and prediabetes[76][87].It is also reduced in certain chronic and inflammatory conditions, including cystic fibrosis[98], reduced PCSK9 expression[109], obstructive sleep apnea[1110] and hidradenitis suppurativa (aka acne inversa)[1211].
A reduced disposition index predicts the evolution of type 2 diabetes[1312]. In a prospective evaluation of the NHANES study, a reduced static disposition index (SPINA-DI) predicted all-cause mortality over 10 years[1413].
Therapeutic measures leading to an improved disposition index include physical activity and nutritional interventions[1514]. In carriers of the risk allele of the gene for the transcription factor TCF7L2 increased dietary fibre intake led to a rising disposition index[1615].