Validation of Doi's weighted average glucose as a measure of post-load glucose excursion for clinical use

Abstract

In this study, we examined the performance of a novel index of glucose excursion (Doi's weighted average glucose [dwAG]) in relation to the conventional measure of area under the oral glucose tolerance test (A-GTT) and the homeostatic model assessment for insulin sensitivity (HOMA-S) and pancreatic beta cell function (HOMA-B). A cross-sectional comparison of the new index was conducted using 66 oral glucose tolerance tests (GTTs) performed at different follow-up times among 27 participants who had undergone surgical subcutaneous fat removal (SSFR). Comparisons across categories were made using box plots and the Kruskal-Wallis one-way ANOVA on ranks. Passing-Bablok regression was used to compare the dwAG against the conventional A-GTT. The Passing-Bablok regression model suggested a cutoff for normality for the A-GTT of 15.14 mmol/L·2h-1 compared to the dwAG's suggested threshold of 6.8 mmol/L. For every 1 mmol/L·2h-1 increase in A-GTT, the dwAG value increased by 0.473 mmol/L. The glucose area under the curve correlated well with the four defined dwAG categories, with at least one of the categories having a different median A-GTT value (KW Chi2 = 52.8 [df = 3], P < 0.001). The HOMA-S tertiles were also associated with significantly different levels of glucose excursion measured through both the dwAG value (KW Chi2 = 11.4 [df = 2], P = 0.003) and A-GTT measure (KW Chi2 = 13.1 [df = 2], P = 0.001). It is concluded that the dwAG value and categories serve as a simple and accurate tool that can be used for interpreting glucose homeostasis across clinical settings.

Figure 1.

Passing–Bablok regression plot showing data and fit for dwAG measure of glucose excursion predicted from A-GTT. Cusum test for linearity, no significant deviation from linearity (P ═ 0.83); Spearman rank correlation coefficient 0.934 (95% CI 0.894–0.959). A-GTT: Area under the glucose tolerance test; dwAG: Doi’s weighted average glucose.

Figure 2.

Association between the A-GTT and dwAG category measure of glucose excursion. A-GTT: Area under the glucose tolerance test; dwAG: Doi’s weighted average glucose.

Figure 3.

Association between glucose excursion (left: Tai’s A-GTT; right: dwAG) and HOMA-S tertile. HOMA-S tertile: Homeostatic Model Assessment for insulin sensitivity tertiles 1, 2, and 3; A-GTT: Area under the glucose tolerance test; dwAG: Doi’s weighted average glucose.

Figure 4.

Relationship of HOMA insulin sensitivity (%) and HOMA beta cell function (%) with the dwAG (mmol/L) demonstrating that both are important in determining the dwAG value. The values of HOMA are centered at 100%, and when both are normal (0%), the dwAG value is about 5 mmol/L. As HOMA insulin sensitivity decreases, the dwAG rises linearly and this is mitigated by an increase in beta cell function which drops the dwAG value. dwAG: Doi’s weighted average glucose; HOMA: Homeostatic model assessment.