Relationship Between Insulin Secretion and Insulin Sensitivity and Its Role in Development of Type 2 Diabetes: Beyond the Disposition Index

Abstract

Hyperbolic relationship between insulin secretion and sensitivityThe relationship between two variables x and y is said to be hyperbolic if 𝑥 × 𝑦 = 𝑐, where c is a constant. The x and y variables can be log-transformed (ln(𝑥 × 𝑦) = ln(𝑥) + ln(𝑦) = ln(𝑐)), and the hyperbolic relationship between these two variables can be re-expressed as linear model such that: ln(𝑦) = ln(𝑐) − 1ln(𝑥), where the regression coefficient in the linear model of ln(𝑦) as a function of ln(𝑥) is 𝛽 = −1. 1,2 Note that β = -1 regardless of the units in which y and x are expressed and regardless of the base of logarithms used. The hypothesis that the relationship between insulin secretion and insulin sensitivity is hyperbolic is usually tested by determining if the slope of a linear model of ln(insulin secretion) as a function of ln(insulin sensitivity) is not significantly different from -1 (i.e., the 95% CI for the slope includes -1). [1][2][3][4] Reasons for using SMA regression to model relationships between insulin secretion and sensitivityWe studied the relationship between insulin secretion and sensitivity by estimating ln(insulin secretion) as a function of ln(insulin sensitivity). The models provided by the SMA regression were chosen after comparing their fit with models from linear and orthogonal regression using standard error of the residuals and confidence intervals for intercepts and slopes, Supplementary Tables S1 and S2. One reason for not considering linear regression models in this case is that this method only accounts for variability in y, while orthogonal and SMA regression models account for variability in both x and y. 5 Also, in our study, linear regression models had the highest standard error of the residuals in all cases, Supplementary Tables S1 and S2. Standard errors of the residuals were the same or slightly lower for orthogonal regression models when compared to those from SMA regression. However, SMA models were chosen because orthogonal regression is most useful when one is studying the relationship between two variables that are estimates of the same entity. 5 Another reason for choosing SMA models was that, their confidence intervals are estimated more efficiently than those for orthogonal regression models, and the confidence intervals for SMA slopes tend to be exact or close to exact in most practical instances. 6 Generally, SMA and orthogonal regression models led to the same conclusions (same qualitative results). In the OGTT dataset, both SMA and orthogonal regression indicated that the slope for ln(CIR120) as a function of ln(ISI0) was essentially -1 in the whole dataset (SMA: (95% CI -0•999 to -0•908); orthogonal: (95% CI -0•996 to -0•686)), significantly different from -1 in NGR subjects only (SMA: (95% CI -0•948 to -0•854); orthogonal: (95% CI -0•869 to -0•656)), and that it was not significantly different from -1 in the IGR subjects only (SMA: (95% CI -1•093 to -0•918); orthogonal: (95% CI -1•180 to -0•853)). In the IV/CLAMP dataset, both approaches suggested that the slope for ln(AIR) as a funct...

Effects of different doses of exercise and diet-induced weight loss on beta-cell function in type 2 diabetes (DOSE-EX): a randomized clinical trial

“…Nevertheless, exercise-induced decreases in insulin secretion during concomitant increases in insulin sensitivity are consistently found in individuals with prediabetes and/or obesity, as well as in healthy people 12 , 23 , 28 . Given that insulin secretion physiologically counterbalances insulin sensitivity as a homeostatic response 15 , 31 , an exercise-induced increase in insulin sensitivity added to diet-induced weight loss may reduce the demand on beta cells, offering beta-cell rest and therefore preserving beta-cell health.…”

“…Sixth, we assessed dietary adherence using self-reporting, which may include information bias 40 . Seventh, although DI is considered the most accurate assessment of beta-cell function 15 , the relationship between insulin secretion and insulin sensitivity is not consistently hyperbolic across levels of glucose tolerance, BMI or measurements of DI 31 . Moreover, an increased beta-cell DI does not necessarily imply improved beta-cell health.…”

“…Moreover, an increased beta-cell DI does not necessarily imply improved beta-cell health. This is because an increase in DI via an increased demand of insulin secretion to compensate for decreased insulin sensitivity (that is, higher allostatic load) has been associated with deterioration of beta-cell function compared with increasing DI via improved insulin sensitivity 31 , 41 , 42 . Therefore, we evaluated both insulin secretion and insulin sensitivity for calculating beta-cell DI, and our results are in line with preclinical and clinical studies suggesting that increasing beta-cell DI through insulin sensitivity is beneficial for beta-cell health.…”

Effects of theABCC8R1420H loss-of-function variant on beta-cell function, diabetes incidence, and retinopathy

“…Insulin secretion compensation and insulin secretion demand were estimated in participants who had at least one measurement of ISI 0 and CIR 120 , using standardized major axis regression9 (see online supplemental material). Estimates of insulin secretion compensation (distances from the regression line) quantify the ability of insulin secretion to compensate for insulin sensitivity and represent a measure of homeostatic response to glucose 8 9.…”

“…They also had higher birth weights which suggests possible exposure to insulin (due to oversecretion) in utero 6. In this population, beta-cell function studied through the relationship between insulin secretion and sensitivity declines as diabetes progresses 7–9. Measures of insulin secretion compensation and insulin secretion demand estimated through the relationship between insulin secretion and sensitivity have also been studied 9…”