Reply to Comment on: Gastaldelli A, Ferrannini E, Miyazaki Y, Matsuda M, DeFronzo RA (2004) Beta cell dysfunction and glucose intolerance: results from the San Antonio metabolism (SAM) study. Diabetologia 43:31–39
KeywordsBeta Cell Plasma Insulin Beta Cell Function Disposition Index Glucose Increment
To the Editor: Dr Jenkins and colleagues raise two issues. The first relates to the validity of the strong inverse relationship between disposition indices and plasma glucose concentrations. The second relates to the effects of obesity on indices of beta cell function. They state that our conclusion appears inconsistent with their finding that “baseline adiposity predicts reduced beta cell function at 6-year follow up, and does so independently of diabetes risk status.” They also state that our findings “are based on spurious correlations.” We offer the following comments.
Firstly, the disposition indices we used are: (ΔI 30/ΔG 30·1/IR) and (ΔI 0–120/ΔG 0–120·1/IR) respectively for the 30-min values of plasma insulin and glucose increments and the average 0 to 120-min values of plasma insulin and glucose increments (see Fig. 3 of original article). The insulin resistance index IR is the inverse of the ratio of total glucose disposal (TGD) to the steady-state plasma insulin concentration (I) achieved during the euglycaemic insulin clamp.
In general, Dr Jenkins develops equations to argue that any disposition index e.g. (ΔI 0–120/ΔG 0–120·1/IR) must be inversely related to plasma glucose (G) through R in (rate of glucose input into the pool). In principle, we do not disagree with Dr Jenkins’ model. We do, however, feel it is grossly oversimplified. TGD is not simply proportional to circulating insulin concentration (I) and is not fully independent of G. Likewise, I is not simply proportional to G, and it is not true even by approximation that quasi-steady-state conditions obtain 2 hours after glucose ingestion. More importantly, insulin sensitivity and, even more importantly, beta cell function cannot be reduced to a single universal index that is the same whatever in vivo test and formula are used to compute it and that applies to all conditions (like IR and β in Dr Jenkins’ model). It cannot therefore be taken for granted that for a given pair of indices of beta cell function and insulin resistance the resulting disposition index should be inversely related to glucose concentrations, with no differences between groups.
A strong correlation would be obvious only by presupposing the existence of universal indices of beta cell function and insulin resistance. In fact, if from our data we were to calculate a disposition index in which the secretion index is the ratio of fasting plasma insulin to fasting plasma glucose (a HOMA-β equivalent), the plot of this disposition index against 2-hour plasma glucose levels would look quite different from those in Figure 3 of our paper . It therefore makes sense to show, as we did, that the relationship between specific disposition indices and glucose tolerance is similar in lean and obese subjects, and to discuss the physiological meaning of the finding. The potential for differences is not only in R in (rate of glucose input into the pool), as Dr Jenkins suggests, but also in factors that differentiate the various aspects of beta cell function, since beta cell function cannot be reduced to a single universal index . For this reason, we maintain that our indices of beta cell function and insulin resistance are indeed close surrogates of the determinants of glucose tolerance according to Dr Jenkins’ model. This, however, is a conclusion and not a premise. On the other hand, we also agree that caution must be exercised when using indices that may exhibit self-correlation.
Our second comment is to point out that in our paper we concluded the following: “Irrespective of the degree of glucose tolerance, when one relates the insulin secretory response to the concomitant severity of insulin resistance, it can be seen that for any given glucose stimulus the beta cell response in obese individuals is superimposable on that of lean subjects.” The cross-sectional data upon which our analyses were based preclude any inference as to whether initially normal beta cell secretory responses may with time deteriorate faster in obese than in non-obese subjects. The paper cited by Jenkins and co-workers  was based on data from 35 women, in whom beta cell function was estimated by HOMA-β, a simple measure of baseline insulin secretion and a rather distant surrogate of beta cell function. While this is an important issue, we believe it remains to be conclusively settled by detailed analysis of beta cell function in large groups of subjects in prospective evaluation.